Skip to content

Error running .pte model with executor_runner #8923

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
corehalt opened this issue Mar 4, 2025 · 7 comments
Open

Error running .pte model with executor_runner #8923

corehalt opened this issue Mar 4, 2025 · 7 comments
Assignees
@JacobSzwejbka
Labels
module: runtime Issues related to the core runtime and code under runtime/ triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@corehalt
Copy link

corehalt commented Mar 4, 2025 •
edited by pytorch-bot bot
Loading

🐛 Describe the bug

I have exported this model:

https://github.com./corehalt/share/raw/refs/heads/main/yolov8n_runtime_issue.pte

with the following code:

from executorch.exir.passes.constant_prop_pass import constant_prop_pass
from executorch.exir.passes.const_prop_pass import ConstPropPass
from executorch import exir

self.model.eval()
aten_dialect_program = torch.export.export(self.model, (self.im,), strict=True)
torch.export.save(aten_dialect_program, "ep.pt2")
aten_dialect_program = constant_prop_pass(aten_dialect_program)
edge_dialect_program = exir.to_edge_transform_and_lower(aten_dialect_program, transform_passes=[ConstPropPass()])
executorch_program = edge_dialect_program.to_executorch(
    exir.ExecutorchBackendConfig(
        passes=[],
        remove_view_copy=False
    )
)
fpte = self.file.with_suffix(".pte")
with open(str(fpte), "wb") as fil:
    fil.write(executorch_program.buffer)

Then I tried to run the model with the official C++ executor_runner and but I get the next error:

gdb --args ../build/third_party/executorch/executor_runner --model_path /tmp/yolov8/yolov8n.pte
(gdb) where
#0  main (argc=1, argv=0x7fffffffde28) at /test/third_party/executorch/examples/portable/executor_runner/executor_runner.cpp:244
(gdb) list
239       ET_LOG(Info, "Inputs prepared.");
240
241       // Run the model.
242       for (uint32_t i = 0; i < FLAGS_num_executions; i++) {
243         Error status = method->execute();
244         ET_CHECK_MSG(
245             status == Error::Ok,
246             "Execution of method %s failed with status 0x%" PRIx32,
247             method_name,
248             (uint32_t)status);
(gdb) print status
$1 = executorch::runtime::Error::InvalidArgument
(gdb) print method_name
$4 = 0x5555561e4718 "forward"

With other models and using the same code, the inference runs without problem.
I also wrote another executor based on the official one but I also get the same error there.
Other things I tried is to use strict=False on torch.export() but still it gives me the same error.

For reference, this is the corresponding output of torch.export.save():

https://github.com./corehalt/share/raw/refs/heads/main/yolov8n_runtime_issue.pt2

Versions

Versions of relevant libraries:
[pip3] executorch==0.6.0a0+7103bb3
[pip3] numpy==2.1.1
[pip3] torch==2.7.0.dev20250131+cpu
[pip3] torchao==0.8.0+git11333ba2
[pip3] torchaudio==2.6.0.dev20250131+cpu
[pip3] torchsr==1.0.4
[pip3] torchvision==0.22.0.dev20250131+cpu

cc @JacobSzwejbka
@iseeyuan iseeyuan added module: runtime Issues related to the core runtime and code under runtime/ triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module labels Mar 4, 2025
@github-project-automation github-project-automation bot moved this to To triage in ExecuTorch Core Mar 4, 2025
@iseeyuan
Copy link
Contributor

iseeyuan commented Mar 4, 2025

@corehalt Thanks for reporting the issue! executorch::runtime::Error::InvalidArgument means that input argument does not match with what the model is expecting. Have you checked the input shape, data type etc, that are the same as the input you used to export?

@JacobSzwejbka JacobSzwejbka self-assigned this Mar 4, 2025
@JacobSzwejbka
Copy link
Contributor

JacobSzwejbka commented Mar 4, 2025 •
edited
Loading

Im guessing the failure is a lifted constant is getting flagged as a user input due to a bug in one of the const prop passes (why are you mixing them?). Can you share the results of

print(aten_dialect_program.graph)
print(aten_dialect_program.graph_signature)

and

print(edge_dialect_program.exported_program().graph_module.graph)
print(edge_dialect_program.exported_program().graph_signature)

@JacobSzwejbka
Copy link
Contributor

Also are you comfortable sharing the model definition?

@JacobSzwejbka JacobSzwejbka moved this from To triage to In progress in ExecuTorch Core Mar 4, 2025
@corehalt
Copy link
Author

corehalt commented Mar 5, 2025

Tried the passes because I was trying to get rid of constants, they are not being folded (#8446), I thought it was an issue on the passes but turned out that there are many constants added in the last steps of the conversion to .pte.

Have you checked the input shape, data type etc, that are the same as the input you used to export?
Yes, I mean, it is just a float32 model and the executor example finds the right shape from the meta data of the input.

@JacobSzwejbka sure, here they are:

After first constant folding pass:

print(aten_dialect_program.graph)
print(aten_dialect_program.graph_signature)

Output:

graph():
    %p_model_0_conv_weight : [num_users=1] = placeholder[target=p_model_0_conv_weight]
    %p_model_0_conv_bias : [num_users=1] = placeholder[target=p_model_0_conv_bias]
    %p_model_1_conv_weight : [num_users=1] = placeholder[target=p_model_1_conv_weight]
    %p_model_1_conv_bias : [num_users=1] = placeholder[target=p_model_1_conv_bias]
    %p_model_2_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv1_conv_weight]
    %p_model_2_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv1_conv_bias]
    %p_model_2_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv2_conv_weight]
    %p_model_2_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv2_conv_bias]
    %p_model_2_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_weight]
    %p_model_2_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_bias]
    %p_model_2_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_weight]
    %p_model_2_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_bias]
    %p_model_3_conv_weight : [num_users=1] = placeholder[target=p_model_3_conv_weight]
    %p_model_3_conv_bias : [num_users=1] = placeholder[target=p_model_3_conv_bias]
    %p_model_4_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv1_conv_weight]
    %p_model_4_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv1_conv_bias]
    %p_model_4_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv2_conv_weight]
    %p_model_4_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv2_conv_bias]
    %p_model_4_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_weight]
    %p_model_4_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_bias]
    %p_model_4_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_weight]
    %p_model_4_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_bias]
    %p_model_4_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_weight]
    %p_model_4_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_bias]
    %p_model_4_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_weight]
    %p_model_4_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_bias]
    %p_model_5_conv_weight : [num_users=1] = placeholder[target=p_model_5_conv_weight]
    %p_model_5_conv_bias : [num_users=1] = placeholder[target=p_model_5_conv_bias]
    %p_model_6_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv1_conv_weight]
    %p_model_6_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv1_conv_bias]
    %p_model_6_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv2_conv_weight]
    %p_model_6_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv2_conv_bias]
    %p_model_6_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_weight]
    %p_model_6_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_bias]
    %p_model_6_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_weight]
    %p_model_6_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_bias]
    %p_model_6_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_weight]
    %p_model_6_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_bias]
    %p_model_6_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_weight]
    %p_model_6_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_bias]
    %p_model_7_conv_weight : [num_users=1] = placeholder[target=p_model_7_conv_weight]
    %p_model_7_conv_bias : [num_users=1] = placeholder[target=p_model_7_conv_bias]
    %p_model_8_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv1_conv_weight]
    %p_model_8_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv1_conv_bias]
    %p_model_8_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv2_conv_weight]
    %p_model_8_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv2_conv_bias]
    %p_model_8_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_weight]
    %p_model_8_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_bias]
    %p_model_8_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_weight]
    %p_model_8_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_bias]
    %p_model_9_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv1_conv_weight]
    %p_model_9_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv1_conv_bias]
    %p_model_9_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv2_conv_weight]
    %p_model_9_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv2_conv_bias]
    %p_model_12_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv1_conv_weight]
    %p_model_12_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv1_conv_bias]
    %p_model_12_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv2_conv_weight]
    %p_model_12_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv2_conv_bias]
    %p_model_12_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_weight]
    %p_model_12_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_bias]
    %p_model_12_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_weight]
    %p_model_12_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_bias]
    %p_model_15_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv1_conv_weight]
    %p_model_15_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv1_conv_bias]
    %p_model_15_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv2_conv_weight]
    %p_model_15_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv2_conv_bias]
    %p_model_15_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_weight]
    %p_model_15_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_bias]
    %p_model_15_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_weight]
    %p_model_15_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_bias]
    %p_model_16_conv_weight : [num_users=1] = placeholder[target=p_model_16_conv_weight]
    %p_model_16_conv_bias : [num_users=1] = placeholder[target=p_model_16_conv_bias]
    %p_model_18_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv1_conv_weight]
    %p_model_18_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv1_conv_bias]
    %p_model_18_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv2_conv_weight]
    %p_model_18_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv2_conv_bias]
    %p_model_18_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_weight]
    %p_model_18_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_bias]
    %p_model_18_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_weight]
    %p_model_18_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_bias]
    %p_model_19_conv_weight : [num_users=1] = placeholder[target=p_model_19_conv_weight]
    %p_model_19_conv_bias : [num_users=1] = placeholder[target=p_model_19_conv_bias]
    %p_model_21_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv1_conv_weight]
    %p_model_21_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv1_conv_bias]
    %p_model_21_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv2_conv_weight]
    %p_model_21_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv2_conv_bias]
    %p_model_21_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_weight]
    %p_model_21_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_bias]
    %p_model_21_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_weight]
    %p_model_21_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_bias]
    %p_model_22_cv2_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_weight]
    %p_model_22_cv2_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_bias]
    %p_model_22_cv2_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_weight]
    %p_model_22_cv2_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_bias]
    %p_model_22_cv2_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_weight]
    %p_model_22_cv2_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_bias]
    %p_model_22_cv2_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_weight]
    %p_model_22_cv2_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_bias]
    %p_model_22_cv2_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_weight]
    %p_model_22_cv2_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_bias]
    %p_model_22_cv2_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_weight]
    %p_model_22_cv2_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_bias]
    %p_model_22_cv2_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_weight]
    %p_model_22_cv2_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_bias]
    %p_model_22_cv2_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_weight]
    %p_model_22_cv2_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_bias]
    %p_model_22_cv2_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_weight]
    %p_model_22_cv2_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_bias]
    %p_model_22_cv3_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_weight]
    %p_model_22_cv3_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_bias]
    %p_model_22_cv3_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_weight]
    %p_model_22_cv3_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_bias]
    %p_model_22_cv3_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_weight]
    %p_model_22_cv3_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_bias]
    %p_model_22_cv3_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_weight]
    %p_model_22_cv3_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_bias]
    %p_model_22_cv3_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_weight]
    %p_model_22_cv3_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_bias]
    %p_model_22_cv3_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_weight]
    %p_model_22_cv3_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_bias]
    %p_model_22_cv3_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_weight]
    %p_model_22_cv3_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_bias]
    %p_model_22_cv3_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_weight]
    %p_model_22_cv3_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_bias]
    %p_model_22_cv3_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_weight]
    %p_model_22_cv3_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_bias]
    %p_model_22_dfl_conv_weight : [num_users=1] = placeholder[target=p_model_22_dfl_conv_weight]
    %c_model_22_anchors : [num_users=1] = placeholder[target=c_model_22_anchors]
    %c_model_22_strides : [num_users=1] = placeholder[target=c_model_22_strides]
    %x : [num_users=1] = placeholder[target=x]
    %conv2d : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%x, %p_model_0_conv_weight, %p_model_0_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu_ : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d,), kwargs = {})
    %conv2d_1 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu_, %p_model_1_conv_weight, %p_model_1_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__1 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_1,), kwargs = {})
    %conv2d_2 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__1, %p_model_2_cv1_conv_weight, %p_model_2_cv1_conv_bias), kwargs = {})
    %silu__2 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_2,), kwargs = {})
    %split_with_sizes : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__2, [16, 16], 1), kwargs = {})
    %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes, 0), kwargs = {})
    %getitem_1 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes, 1), kwargs = {})
    %conv2d_3 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_1, %p_model_2_m_0_cv1_conv_weight, %p_model_2_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__3 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_3,), kwargs = {})
    %conv2d_4 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__3, %p_model_2_m_0_cv2_conv_weight, %p_model_2_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__4 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_4,), kwargs = {})
    %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_1, %silu__4), kwargs = {})
    %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem, %getitem_1, %add], 1), kwargs = {})
    %conv2d_5 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat, %p_model_2_cv2_conv_weight, %p_model_2_cv2_conv_bias), kwargs = {})
    %silu__5 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_5,), kwargs = {})
    %conv2d_6 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__5, %p_model_3_conv_weight, %p_model_3_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__6 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_6,), kwargs = {})
    %conv2d_7 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__6, %p_model_4_cv1_conv_weight, %p_model_4_cv1_conv_bias), kwargs = {})
    %silu__7 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_7,), kwargs = {})
    %split_with_sizes_1 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__7, [32, 32], 1), kwargs = {})
    %getitem_2 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_1, 0), kwargs = {})
    %getitem_3 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_1, 1), kwargs = {})
    %conv2d_8 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_3, %p_model_4_m_0_cv1_conv_weight, %p_model_4_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__8 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_8,), kwargs = {})
    %conv2d_9 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__8, %p_model_4_m_0_cv2_conv_weight, %p_model_4_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__9 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_9,), kwargs = {})
    %add_1 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_3, %silu__9), kwargs = {})
    %conv2d_10 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%add_1, %p_model_4_m_1_cv1_conv_weight, %p_model_4_m_1_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__10 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_10,), kwargs = {})
    %conv2d_11 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__10, %p_model_4_m_1_cv2_conv_weight, %p_model_4_m_1_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__11 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_11,), kwargs = {})
    %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %silu__11), kwargs = {})
    %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_2, %getitem_3, %add_1, %add_2], 1), kwargs = {})
    %conv2d_12 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_1, %p_model_4_cv2_conv_weight, %p_model_4_cv2_conv_bias), kwargs = {})
    %silu__12 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_12,), kwargs = {})
    %conv2d_13 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__12, %p_model_5_conv_weight, %p_model_5_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__13 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_13,), kwargs = {})
    %conv2d_14 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__13, %p_model_6_cv1_conv_weight, %p_model_6_cv1_conv_bias), kwargs = {})
    %silu__14 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_14,), kwargs = {})
    %split_with_sizes_2 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__14, [64, 64], 1), kwargs = {})
    %getitem_4 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_2, 0), kwargs = {})
    %getitem_5 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_2, 1), kwargs = {})
    %conv2d_15 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_5, %p_model_6_m_0_cv1_conv_weight, %p_model_6_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__15 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_15,), kwargs = {})
    %conv2d_16 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__15, %p_model_6_m_0_cv2_conv_weight, %p_model_6_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__16 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_16,), kwargs = {})
    %add_3 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_5, %silu__16), kwargs = {})
    %conv2d_17 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%add_3, %p_model_6_m_1_cv1_conv_weight, %p_model_6_m_1_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__17 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_17,), kwargs = {})
    %conv2d_18 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__17, %p_model_6_m_1_cv2_conv_weight, %p_model_6_m_1_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__18 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_18,), kwargs = {})
    %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_3, %silu__18), kwargs = {})
    %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_4, %getitem_5, %add_3, %add_4], 1), kwargs = {})
    %conv2d_19 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_2, %p_model_6_cv2_conv_weight, %p_model_6_cv2_conv_bias), kwargs = {})
    %silu__19 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_19,), kwargs = {})
    %conv2d_20 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__19, %p_model_7_conv_weight, %p_model_7_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__20 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_20,), kwargs = {})
    %conv2d_21 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__20, %p_model_8_cv1_conv_weight, %p_model_8_cv1_conv_bias), kwargs = {})
    %silu__21 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_21,), kwargs = {})
    %split_with_sizes_3 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__21, [128, 128], 1), kwargs = {})
    %getitem_6 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_3, 0), kwargs = {})
    %getitem_7 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_3, 1), kwargs = {})
    %conv2d_22 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_7, %p_model_8_m_0_cv1_conv_weight, %p_model_8_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__22 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_22,), kwargs = {})
    %conv2d_23 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__22, %p_model_8_m_0_cv2_conv_weight, %p_model_8_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__23 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_23,), kwargs = {})
    %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_7, %silu__23), kwargs = {})
    %cat_3 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_6, %getitem_7, %add_5], 1), kwargs = {})
    %conv2d_24 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_3, %p_model_8_cv2_conv_weight, %p_model_8_cv2_conv_bias), kwargs = {})
    %silu__24 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_24,), kwargs = {})
    %conv2d_25 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__24, %p_model_9_cv1_conv_weight, %p_model_9_cv1_conv_bias), kwargs = {})
    %silu__25 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_25,), kwargs = {})
    %max_pool2d : [num_users=2] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%silu__25, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %max_pool2d_1 : [num_users=2] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%max_pool2d, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %max_pool2d_2 : [num_users=1] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%max_pool2d_1, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %cat_4 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__25, %max_pool2d, %max_pool2d_1, %max_pool2d_2], 1), kwargs = {})
    %conv2d_26 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_4, %p_model_9_cv2_conv_weight, %p_model_9_cv2_conv_bias), kwargs = {})
    %silu__26 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_26,), kwargs = {})
    %upsample_nearest2d : [num_users=1] = call_function[target=torch.ops.aten.upsample_nearest2d.vec](args = (%silu__26, None, [2.0, 2.0]), kwargs = {})
    %cat_5 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%upsample_nearest2d, %silu__19], 1), kwargs = {})
    %conv2d_27 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_5, %p_model_12_cv1_conv_weight, %p_model_12_cv1_conv_bias), kwargs = {})
    %silu__27 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_27,), kwargs = {})
    %split_with_sizes_4 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__27, [64, 64], 1), kwargs = {})
    %getitem_8 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_4, 0), kwargs = {})
    %getitem_9 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_4, 1), kwargs = {})
    %conv2d_28 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_9, %p_model_12_m_0_cv1_conv_weight, %p_model_12_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__28 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_28,), kwargs = {})
    %conv2d_29 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__28, %p_model_12_m_0_cv2_conv_weight, %p_model_12_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__29 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_29,), kwargs = {})
    %cat_6 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_8, %getitem_9, %silu__29], 1), kwargs = {})
    %conv2d_30 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_6, %p_model_12_cv2_conv_weight, %p_model_12_cv2_conv_bias), kwargs = {})
    %silu__30 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_30,), kwargs = {})
    %upsample_nearest2d_1 : [num_users=1] = call_function[target=torch.ops.aten.upsample_nearest2d.vec](args = (%silu__30, None, [2.0, 2.0]), kwargs = {})
    %cat_7 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%upsample_nearest2d_1, %silu__12], 1), kwargs = {})
    %conv2d_31 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_7, %p_model_15_cv1_conv_weight, %p_model_15_cv1_conv_bias), kwargs = {})
    %silu__31 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_31,), kwargs = {})
    %split_with_sizes_5 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__31, [32, 32], 1), kwargs = {})
    %getitem_10 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_5, 0), kwargs = {})
    %getitem_11 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_5, 1), kwargs = {})
    %conv2d_32 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_11, %p_model_15_m_0_cv1_conv_weight, %p_model_15_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__32 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_32,), kwargs = {})
    %conv2d_33 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__32, %p_model_15_m_0_cv2_conv_weight, %p_model_15_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__33 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_33,), kwargs = {})
    %cat_8 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_10, %getitem_11, %silu__33], 1), kwargs = {})
    %conv2d_34 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_8, %p_model_15_cv2_conv_weight, %p_model_15_cv2_conv_bias), kwargs = {})
    %silu__34 : [num_users=3] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_34,), kwargs = {})
    %conv2d_35 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_16_conv_weight, %p_model_16_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__35 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_35,), kwargs = {})
    %cat_9 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__35, %silu__30], 1), kwargs = {})
    %conv2d_36 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_9, %p_model_18_cv1_conv_weight, %p_model_18_cv1_conv_bias), kwargs = {})
    %silu__36 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_36,), kwargs = {})
    %split_with_sizes_6 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__36, [64, 64], 1), kwargs = {})
    %getitem_12 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_6, 0), kwargs = {})
    %getitem_13 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_6, 1), kwargs = {})
    %conv2d_37 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_13, %p_model_18_m_0_cv1_conv_weight, %p_model_18_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__37 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_37,), kwargs = {})
    %conv2d_38 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__37, %p_model_18_m_0_cv2_conv_weight, %p_model_18_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__38 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_38,), kwargs = {})
    %cat_10 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_12, %getitem_13, %silu__38], 1), kwargs = {})
    %conv2d_39 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_10, %p_model_18_cv2_conv_weight, %p_model_18_cv2_conv_bias), kwargs = {})
    %silu__39 : [num_users=3] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_39,), kwargs = {})
    %conv2d_40 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_19_conv_weight, %p_model_19_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__40 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_40,), kwargs = {})
    %cat_11 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__40, %silu__26], 1), kwargs = {})
    %conv2d_41 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_11, %p_model_21_cv1_conv_weight, %p_model_21_cv1_conv_bias), kwargs = {})
    %silu__41 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_41,), kwargs = {})
    %split_with_sizes_7 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__41, [128, 128], 1), kwargs = {})
    %getitem_14 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_7, 0), kwargs = {})
    %getitem_15 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_7, 1), kwargs = {})
    %conv2d_42 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_15, %p_model_21_m_0_cv1_conv_weight, %p_model_21_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__42 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_42,), kwargs = {})
    %conv2d_43 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__42, %p_model_21_m_0_cv2_conv_weight, %p_model_21_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__43 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_43,), kwargs = {})
    %cat_12 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_14, %getitem_15, %silu__43], 1), kwargs = {})
    %conv2d_44 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_12, %p_model_21_cv2_conv_weight, %p_model_21_cv2_conv_bias), kwargs = {})
    %silu__44 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_44,), kwargs = {})
    %conv2d_45 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_22_cv2_0_0_conv_weight, %p_model_22_cv2_0_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__45 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_45,), kwargs = {})
    %conv2d_46 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__45, %p_model_22_cv2_0_1_conv_weight, %p_model_22_cv2_0_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__46 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_46,), kwargs = {})
    %conv2d_47 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__46, %p_model_22_cv2_0_2_weight, %p_model_22_cv2_0_2_bias), kwargs = {})
    %conv2d_48 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_22_cv3_0_0_conv_weight, %p_model_22_cv3_0_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__47 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_48,), kwargs = {})
    %conv2d_49 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__47, %p_model_22_cv3_0_1_conv_weight, %p_model_22_cv3_0_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__48 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_49,), kwargs = {})
    %conv2d_50 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__48, %p_model_22_cv3_0_2_weight, %p_model_22_cv3_0_2_bias), kwargs = {})
    %cat_13 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_47, %conv2d_50], 1), kwargs = {})
    %conv2d_51 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_22_cv2_1_0_conv_weight, %p_model_22_cv2_1_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__49 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_51,), kwargs = {})
    %conv2d_52 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__49, %p_model_22_cv2_1_1_conv_weight, %p_model_22_cv2_1_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__50 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_52,), kwargs = {})
    %conv2d_53 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__50, %p_model_22_cv2_1_2_weight, %p_model_22_cv2_1_2_bias), kwargs = {})
    %conv2d_54 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_22_cv3_1_0_conv_weight, %p_model_22_cv3_1_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__51 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_54,), kwargs = {})
    %conv2d_55 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__51, %p_model_22_cv3_1_1_conv_weight, %p_model_22_cv3_1_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__52 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_55,), kwargs = {})
    %conv2d_56 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__52, %p_model_22_cv3_1_2_weight, %p_model_22_cv3_1_2_bias), kwargs = {})
    %cat_14 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_53, %conv2d_56], 1), kwargs = {})
    %conv2d_57 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__44, %p_model_22_cv2_2_0_conv_weight, %p_model_22_cv2_2_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__53 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_57,), kwargs = {})
    %conv2d_58 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__53, %p_model_22_cv2_2_1_conv_weight, %p_model_22_cv2_2_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__54 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_58,), kwargs = {})
    %conv2d_59 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__54, %p_model_22_cv2_2_2_weight, %p_model_22_cv2_2_2_bias), kwargs = {})
    %conv2d_60 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__44, %p_model_22_cv3_2_0_conv_weight, %p_model_22_cv3_2_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__55 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_60,), kwargs = {})
    %conv2d_61 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__55, %p_model_22_cv3_2_1_conv_weight, %p_model_22_cv3_2_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__56 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_61,), kwargs = {})
    %conv2d_62 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__56, %p_model_22_cv3_2_2_weight, %p_model_22_cv3_2_2_bias), kwargs = {})
    %cat_15 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_59, %conv2d_62], 1), kwargs = {})
    %view : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_13, [1, 144, -1]), kwargs = {})
    %view_1 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_14, [1, 144, -1]), kwargs = {})
    %view_2 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_15, [1, 144, -1]), kwargs = {})
    %cat_16 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%view, %view_1, %view_2], 2), kwargs = {})
    %split_with_sizes_8 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%cat_16, [64, 80], 1), kwargs = {})
    %getitem_16 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_8, 0), kwargs = {})
    %getitem_17 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_8, 1), kwargs = {})
    %view_3 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%getitem_16, [1, 4, 16, 8400]), kwargs = {})
    %transpose : [num_users=1] = call_function[target=torch.ops.aten.transpose.int](args = (%view_3, 2, 1), kwargs = {})
    %softmax : [num_users=1] = call_function[target=torch.ops.aten.softmax.int](args = (%transpose, 1), kwargs = {})
    %conv2d_63 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%softmax, %p_model_22_dfl_conv_weight), kwargs = {})
    %view_4 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%conv2d_63, [1, 4, 8400]), kwargs = {})
    %unsqueeze : [num_users=2] = call_function[target=torch.ops.aten.unsqueeze.default](args = (%c_model_22_anchors, 0), kwargs = {})
    %chunk : [num_users=2] = call_function[target=torch.ops.aten.chunk.default](args = (%view_4, 2, 1), kwargs = {})
    %getitem_18 : [num_users=1] = call_function[target=operator.getitem](args = (%chunk, 0), kwargs = {})
    %getitem_19 : [num_users=1] = call_function[target=operator.getitem](args = (%chunk, 1), kwargs = {})
    %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%unsqueeze, %getitem_18), kwargs = {})
    %add_6 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%unsqueeze, %getitem_19), kwargs = {})
    %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub, %add_6), kwargs = {})
    %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_7, 2), kwargs = {})
    %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_6, %sub), kwargs = {})
    %cat_17 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%div, %sub_1], 1), kwargs = {})
    %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%cat_17, %c_model_22_strides), kwargs = {})
    %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_17,), kwargs = {})
    %cat_18 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mul, %sigmoid], 1), kwargs = {})
    return (cat_18,)
ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_weight'), target='model.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_bias'), target='model.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_weight'), target='model.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_bias'), target='model.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_weight'), target='model.2.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_bias'), target='model.2.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_weight'), target='model.2.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_bias'), target='model.2.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_weight'), target='model.2.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_bias'), target='model.2.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_weight'), target='model.2.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_bias'), target='model.2.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_weight'), target='model.3.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_bias'), target='model.3.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_weight'), target='model.4.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_bias'), target='model.4.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_weight'), target='model.4.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_bias'), target='model.4.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_weight'), target='model.4.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_bias'), target='model.4.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_weight'), target='model.4.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_bias'), target='model.4.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_weight'), target='model.4.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_bias'), target='model.4.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_weight'), target='model.4.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_bias'), target='model.4.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_weight'), target='model.5.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_bias'), target='model.5.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_weight'), target='model.6.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_bias'), target='model.6.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_weight'), target='model.6.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_bias'), target='model.6.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_weight'), target='model.6.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_bias'), target='model.6.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_weight'), target='model.6.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_bias'), target='model.6.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_weight'), target='model.6.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_bias'), target='model.6.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_weight'), target='model.6.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_bias'), target='model.6.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_weight'), target='model.7.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_bias'), target='model.7.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_weight'), target='model.8.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_bias'), target='model.8.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_weight'), target='model.8.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_bias'), target='model.8.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_weight'), target='model.8.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_bias'), target='model.8.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_weight'), target='model.8.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_bias'), target='model.8.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_weight'), target='model.9.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_bias'), target='model.9.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_weight'), target='model.9.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_bias'), target='model.9.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_weight'), target='model.12.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_bias'), target='model.12.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_weight'), target='model.12.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_bias'), target='model.12.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_weight'), target='model.12.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_bias'), target='model.12.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_weight'), target='model.12.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_bias'), target='model.12.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_weight'), target='model.15.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_bias'), target='model.15.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_weight'), target='model.15.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_bias'), target='model.15.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_weight'), target='model.15.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_bias'), target='model.15.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_weight'), target='model.15.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_bias'), target='model.15.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_weight'), target='model.16.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_bias'), target='model.16.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_weight'), target='model.18.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_bias'), target='model.18.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_weight'), target='model.18.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_bias'), target='model.18.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_weight'), target='model.18.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_bias'), target='model.18.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_weight'), target='model.18.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_bias'), target='model.18.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_weight'), target='model.19.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_bias'), target='model.19.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_weight'), target='model.21.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_bias'), target='model.21.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_weight'), target='model.21.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_bias'), target='model.21.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_weight'), target='model.21.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_bias'), target='model.21.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_weight'), target='model.21.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_bias'), target='model.21.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_weight'), target='model.22.cv2.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_bias'), target='model.22.cv2.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_weight'), target='model.22.cv2.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_bias'), target='model.22.cv2.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_weight'), target='model.22.cv2.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_bias'), target='model.22.cv2.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_weight'), target='model.22.cv2.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_bias'), target='model.22.cv2.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_weight'), target='model.22.cv2.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_bias'), target='model.22.cv2.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_weight'), target='model.22.cv2.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_bias'), target='model.22.cv2.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_weight'), target='model.22.cv2.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_bias'), target='model.22.cv2.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_weight'), target='model.22.cv2.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_bias'), target='model.22.cv2.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_weight'), target='model.22.cv2.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_bias'), target='model.22.cv2.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_weight'), target='model.22.cv3.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_bias'), target='model.22.cv3.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_weight'), target='model.22.cv3.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_bias'), target='model.22.cv3.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_weight'), target='model.22.cv3.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_bias'), target='model.22.cv3.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_weight'), target='model.22.cv3.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_bias'), target='model.22.cv3.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_weight'), target='model.22.cv3.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_bias'), target='model.22.cv3.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_weight'), target='model.22.cv3.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_bias'), target='model.22.cv3.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_weight'), target='model.22.cv3.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_bias'), target='model.22.cv3.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_weight'), target='model.22.cv3.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_bias'), target='model.22.cv3.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_weight'), target='model.22.cv3.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_bias'), target='model.22.cv3.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_dfl_conv_weight'), target='model.22.dfl.conv.weight', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_anchors'), target='model.22.anchors', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_strides'), target='model.22.strides', persistent=None), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='x'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='cat_18'), target=None)])


----

ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_weight'), target='model.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_bias'), target='model.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_weight'), target='model.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_bias'), target='model.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_weight'), target='model.2.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_bias'), target='model.2.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_weight'), target='model.2.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_bias'), target='model.2.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_weight'), target='model.2.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_bias'), target='model.2.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_weight'), target='model.2.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_bias'), target='model.2.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_weight'), target='model.3.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_bias'), target='model.3.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_weight'), target='model.4.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_bias'), target='model.4.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_weight'), target='model.4.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_bias'), target='model.4.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_weight'), target='model.4.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_bias'), target='model.4.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_weight'), target='model.4.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_bias'), target='model.4.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_weight'), target='model.4.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_bias'), target='model.4.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_weight'), target='model.4.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_bias'), target='model.4.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_weight'), target='model.5.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_bias'), target='model.5.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_weight'), target='model.6.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_bias'), target='model.6.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_weight'), target='model.6.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_bias'), target='model.6.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_weight'), target='model.6.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_bias'), target='model.6.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_weight'), target='model.6.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_bias'), target='model.6.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_weight'), target='model.6.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_bias'), target='model.6.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_weight'), target='model.6.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_bias'), target='model.6.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_weight'), target='model.7.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_bias'), target='model.7.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_weight'), target='model.8.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_bias'), target='model.8.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_weight'), target='model.8.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_bias'), target='model.8.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_weight'), target='model.8.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_bias'), target='model.8.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_weight'), target='model.8.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_bias'), target='model.8.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_weight'), target='model.9.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_bias'), target='model.9.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_weight'), target='model.9.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_bias'), target='model.9.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_weight'), target='model.12.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_bias'), target='model.12.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_weight'), target='model.12.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_bias'), target='model.12.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_weight'), target='model.12.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_bias'), target='model.12.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_weight'), target='model.12.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_bias'), target='model.12.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_weight'), target='model.15.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_bias'), target='model.15.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_weight'), target='model.15.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_bias'), target='model.15.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_weight'), target='model.15.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_bias'), target='model.15.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_weight'), target='model.15.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_bias'), target='model.15.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_weight'), target='model.16.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_bias'), target='model.16.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_weight'), target='model.18.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_bias'), target='model.18.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_weight'), target='model.18.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_bias'), target='model.18.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_weight'), target='model.18.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_bias'), target='model.18.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_weight'), target='model.18.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_bias'), target='model.18.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_weight'), target='model.19.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_bias'), target='model.19.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_weight'), target='model.21.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_bias'), target='model.21.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_weight'), target='model.21.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_bias'), target='model.21.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_weight'), target='model.21.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_bias'), target='model.21.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_weight'), target='model.21.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_bias'), target='model.21.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_weight'), target='model.22.cv2.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_bias'), target='model.22.cv2.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_weight'), target='model.22.cv2.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_bias'), target='model.22.cv2.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_weight'), target='model.22.cv2.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_bias'), target='model.22.cv2.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_weight'), target='model.22.cv2.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_bias'), target='model.22.cv2.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_weight'), target='model.22.cv2.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_bias'), target='model.22.cv2.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_weight'), target='model.22.cv2.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_bias'), target='model.22.cv2.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_weight'), target='model.22.cv2.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_bias'), target='model.22.cv2.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_weight'), target='model.22.cv2.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_bias'), target='model.22.cv2.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_weight'), target='model.22.cv2.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_bias'), target='model.22.cv2.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_weight'), target='model.22.cv3.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_bias'), target='model.22.cv3.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_weight'), target='model.22.cv3.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_bias'), target='model.22.cv3.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_weight'), target='model.22.cv3.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_bias'), target='model.22.cv3.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_weight'), target='model.22.cv3.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_bias'), target='model.22.cv3.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_weight'), target='model.22.cv3.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_bias'), target='model.22.cv3.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_weight'), target='model.22.cv3.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_bias'), target='model.22.cv3.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_weight'), target='model.22.cv3.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_bias'), target='model.22.cv3.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_weight'), target='model.22.cv3.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_bias'), target='model.22.cv3.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_weight'), target='model.22.cv3.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_bias'), target='model.22.cv3.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_dfl_conv_weight'), target='model.22.dfl.conv.weight', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_anchors'), target='model.22.anchors', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_strides'), target='model.22.strides', persistent=None), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='x'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='cat_18'), target=None)])

After second constant folding pass:

print(aten_dialect_program.graph)
print(aten_dialect_program.graph_signature)

Output:

graph():
    %p_model_0_conv_weight : [num_users=1] = placeholder[target=p_model_0_conv_weight]
    %p_model_0_conv_bias : [num_users=1] = placeholder[target=p_model_0_conv_bias]
    %p_model_1_conv_weight : [num_users=1] = placeholder[target=p_model_1_conv_weight]
    %p_model_1_conv_bias : [num_users=1] = placeholder[target=p_model_1_conv_bias]
    %p_model_2_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv1_conv_weight]
    %p_model_2_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv1_conv_bias]
    %p_model_2_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv2_conv_weight]
    %p_model_2_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv2_conv_bias]
    %p_model_2_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_weight]
    %p_model_2_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_bias]
    %p_model_2_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_weight]
    %p_model_2_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_bias]
    %p_model_3_conv_weight : [num_users=1] = placeholder[target=p_model_3_conv_weight]
    %p_model_3_conv_bias : [num_users=1] = placeholder[target=p_model_3_conv_bias]
    %p_model_4_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv1_conv_weight]
    %p_model_4_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv1_conv_bias]
    %p_model_4_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv2_conv_weight]
    %p_model_4_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv2_conv_bias]
    %p_model_4_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_weight]
    %p_model_4_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_bias]
    %p_model_4_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_weight]
    %p_model_4_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_bias]
    %p_model_4_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_weight]
    %p_model_4_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_bias]
    %p_model_4_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_weight]
    %p_model_4_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_bias]
    %p_model_5_conv_weight : [num_users=1] = placeholder[target=p_model_5_conv_weight]
    %p_model_5_conv_bias : [num_users=1] = placeholder[target=p_model_5_conv_bias]
    %p_model_6_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv1_conv_weight]
    %p_model_6_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv1_conv_bias]
    %p_model_6_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv2_conv_weight]
    %p_model_6_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv2_conv_bias]
    %p_model_6_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_weight]
    %p_model_6_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_bias]
    %p_model_6_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_weight]
    %p_model_6_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_bias]
    %p_model_6_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_weight]
    %p_model_6_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_bias]
    %p_model_6_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_weight]
    %p_model_6_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_bias]
    %p_model_7_conv_weight : [num_users=1] = placeholder[target=p_model_7_conv_weight]
    %p_model_7_conv_bias : [num_users=1] = placeholder[target=p_model_7_conv_bias]
    %p_model_8_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv1_conv_weight]
    %p_model_8_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv1_conv_bias]
    %p_model_8_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv2_conv_weight]
    %p_model_8_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv2_conv_bias]
    %p_model_8_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_weight]
    %p_model_8_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_bias]
    %p_model_8_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_weight]
    %p_model_8_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_bias]
    %p_model_9_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv1_conv_weight]
    %p_model_9_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv1_conv_bias]
    %p_model_9_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv2_conv_weight]
    %p_model_9_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv2_conv_bias]
    %p_model_12_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv1_conv_weight]
    %p_model_12_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv1_conv_bias]
    %p_model_12_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv2_conv_weight]
    %p_model_12_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv2_conv_bias]
    %p_model_12_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_weight]
    %p_model_12_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_bias]
    %p_model_12_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_weight]
    %p_model_12_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_bias]
    %p_model_15_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv1_conv_weight]
    %p_model_15_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv1_conv_bias]
    %p_model_15_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv2_conv_weight]
    %p_model_15_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv2_conv_bias]
    %p_model_15_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_weight]
    %p_model_15_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_bias]
    %p_model_15_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_weight]
    %p_model_15_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_bias]
    %p_model_16_conv_weight : [num_users=1] = placeholder[target=p_model_16_conv_weight]
    %p_model_16_conv_bias : [num_users=1] = placeholder[target=p_model_16_conv_bias]
    %p_model_18_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv1_conv_weight]
    %p_model_18_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv1_conv_bias]
    %p_model_18_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv2_conv_weight]
    %p_model_18_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv2_conv_bias]
    %p_model_18_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_weight]
    %p_model_18_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_bias]
    %p_model_18_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_weight]
    %p_model_18_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_bias]
    %p_model_19_conv_weight : [num_users=1] = placeholder[target=p_model_19_conv_weight]
    %p_model_19_conv_bias : [num_users=1] = placeholder[target=p_model_19_conv_bias]
    %p_model_21_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv1_conv_weight]
    %p_model_21_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv1_conv_bias]
    %p_model_21_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv2_conv_weight]
    %p_model_21_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv2_conv_bias]
    %p_model_21_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_weight]
    %p_model_21_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_bias]
    %p_model_21_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_weight]
    %p_model_21_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_bias]
    %p_model_22_cv2_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_weight]
    %p_model_22_cv2_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_bias]
    %p_model_22_cv2_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_weight]
    %p_model_22_cv2_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_bias]
    %p_model_22_cv2_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_weight]
    %p_model_22_cv2_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_bias]
    %p_model_22_cv2_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_weight]
    %p_model_22_cv2_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_bias]
    %p_model_22_cv2_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_weight]
    %p_model_22_cv2_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_bias]
    %p_model_22_cv2_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_weight]
    %p_model_22_cv2_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_bias]
    %p_model_22_cv2_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_weight]
    %p_model_22_cv2_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_bias]
    %p_model_22_cv2_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_weight]
    %p_model_22_cv2_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_bias]
    %p_model_22_cv2_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_weight]
    %p_model_22_cv2_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_bias]
    %p_model_22_cv3_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_weight]
    %p_model_22_cv3_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_bias]
    %p_model_22_cv3_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_weight]
    %p_model_22_cv3_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_bias]
    %p_model_22_cv3_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_weight]
    %p_model_22_cv3_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_bias]
    %p_model_22_cv3_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_weight]
    %p_model_22_cv3_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_bias]
    %p_model_22_cv3_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_weight]
    %p_model_22_cv3_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_bias]
    %p_model_22_cv3_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_weight]
    %p_model_22_cv3_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_bias]
    %p_model_22_cv3_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_weight]
    %p_model_22_cv3_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_bias]
    %p_model_22_cv3_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_weight]
    %p_model_22_cv3_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_bias]
    %p_model_22_cv3_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_weight]
    %p_model_22_cv3_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_bias]
    %p_model_22_dfl_conv_weight : [num_users=1] = placeholder[target=p_model_22_dfl_conv_weight]
    %c_model_22_strides : [num_users=1] = placeholder[target=c_model_22_strides]
    %_prop_tensor_constant2 : [num_users=2] = placeholder[target=_prop_tensor_constant2]
    %x : [num_users=1] = placeholder[target=x]
    %conv2d : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%x, %p_model_0_conv_weight, %p_model_0_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu_ : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d,), kwargs = {})
    %conv2d_1 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu_, %p_model_1_conv_weight, %p_model_1_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__1 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_1,), kwargs = {})
    %conv2d_2 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__1, %p_model_2_cv1_conv_weight, %p_model_2_cv1_conv_bias), kwargs = {})
    %silu__2 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_2,), kwargs = {})
    %split_with_sizes : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__2, [16, 16], 1), kwargs = {})
    %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes, 0), kwargs = {})
    %getitem_1 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes, 1), kwargs = {})
    %conv2d_3 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_1, %p_model_2_m_0_cv1_conv_weight, %p_model_2_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__3 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_3,), kwargs = {})
    %conv2d_4 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__3, %p_model_2_m_0_cv2_conv_weight, %p_model_2_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__4 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_4,), kwargs = {})
    %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_1, %silu__4), kwargs = {})
    %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem, %getitem_1, %add], 1), kwargs = {})
    %conv2d_5 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat, %p_model_2_cv2_conv_weight, %p_model_2_cv2_conv_bias), kwargs = {})
    %silu__5 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_5,), kwargs = {})
    %conv2d_6 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__5, %p_model_3_conv_weight, %p_model_3_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__6 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_6,), kwargs = {})
    %conv2d_7 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__6, %p_model_4_cv1_conv_weight, %p_model_4_cv1_conv_bias), kwargs = {})
    %silu__7 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_7,), kwargs = {})
    %split_with_sizes_1 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__7, [32, 32], 1), kwargs = {})
    %getitem_2 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_1, 0), kwargs = {})
    %getitem_3 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_1, 1), kwargs = {})
    %conv2d_8 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_3, %p_model_4_m_0_cv1_conv_weight, %p_model_4_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__8 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_8,), kwargs = {})
    %conv2d_9 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__8, %p_model_4_m_0_cv2_conv_weight, %p_model_4_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__9 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_9,), kwargs = {})
    %add_1 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_3, %silu__9), kwargs = {})
    %conv2d_10 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%add_1, %p_model_4_m_1_cv1_conv_weight, %p_model_4_m_1_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__10 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_10,), kwargs = {})
    %conv2d_11 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__10, %p_model_4_m_1_cv2_conv_weight, %p_model_4_m_1_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__11 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_11,), kwargs = {})
    %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %silu__11), kwargs = {})
    %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_2, %getitem_3, %add_1, %add_2], 1), kwargs = {})
    %conv2d_12 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_1, %p_model_4_cv2_conv_weight, %p_model_4_cv2_conv_bias), kwargs = {})
    %silu__12 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_12,), kwargs = {})
    %conv2d_13 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__12, %p_model_5_conv_weight, %p_model_5_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__13 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_13,), kwargs = {})
    %conv2d_14 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__13, %p_model_6_cv1_conv_weight, %p_model_6_cv1_conv_bias), kwargs = {})
    %silu__14 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_14,), kwargs = {})
    %split_with_sizes_2 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__14, [64, 64], 1), kwargs = {})
    %getitem_4 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_2, 0), kwargs = {})
    %getitem_5 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_2, 1), kwargs = {})
    %conv2d_15 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_5, %p_model_6_m_0_cv1_conv_weight, %p_model_6_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__15 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_15,), kwargs = {})
    %conv2d_16 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__15, %p_model_6_m_0_cv2_conv_weight, %p_model_6_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__16 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_16,), kwargs = {})
    %add_3 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_5, %silu__16), kwargs = {})
    %conv2d_17 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%add_3, %p_model_6_m_1_cv1_conv_weight, %p_model_6_m_1_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__17 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_17,), kwargs = {})
    %conv2d_18 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__17, %p_model_6_m_1_cv2_conv_weight, %p_model_6_m_1_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__18 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_18,), kwargs = {})
    %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_3, %silu__18), kwargs = {})
    %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_4, %getitem_5, %add_3, %add_4], 1), kwargs = {})
    %conv2d_19 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_2, %p_model_6_cv2_conv_weight, %p_model_6_cv2_conv_bias), kwargs = {})
    %silu__19 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_19,), kwargs = {})
    %conv2d_20 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__19, %p_model_7_conv_weight, %p_model_7_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__20 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_20,), kwargs = {})
    %conv2d_21 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__20, %p_model_8_cv1_conv_weight, %p_model_8_cv1_conv_bias), kwargs = {})
    %silu__21 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_21,), kwargs = {})
    %split_with_sizes_3 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__21, [128, 128], 1), kwargs = {})
    %getitem_6 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_3, 0), kwargs = {})
    %getitem_7 : [num_users=3] = call_function[target=operator.getitem](args = (%split_with_sizes_3, 1), kwargs = {})
    %conv2d_22 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_7, %p_model_8_m_0_cv1_conv_weight, %p_model_8_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__22 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_22,), kwargs = {})
    %conv2d_23 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__22, %p_model_8_m_0_cv2_conv_weight, %p_model_8_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__23 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_23,), kwargs = {})
    %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_7, %silu__23), kwargs = {})
    %cat_3 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_6, %getitem_7, %add_5], 1), kwargs = {})
    %conv2d_24 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_3, %p_model_8_cv2_conv_weight, %p_model_8_cv2_conv_bias), kwargs = {})
    %silu__24 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_24,), kwargs = {})
    %conv2d_25 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__24, %p_model_9_cv1_conv_weight, %p_model_9_cv1_conv_bias), kwargs = {})
    %silu__25 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_25,), kwargs = {})
    %max_pool2d : [num_users=2] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%silu__25, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %max_pool2d_1 : [num_users=2] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%max_pool2d, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %max_pool2d_2 : [num_users=1] = call_function[target=torch.ops.aten.max_pool2d.default](args = (%max_pool2d_1, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %cat_4 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__25, %max_pool2d, %max_pool2d_1, %max_pool2d_2], 1), kwargs = {})
    %conv2d_26 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_4, %p_model_9_cv2_conv_weight, %p_model_9_cv2_conv_bias), kwargs = {})
    %silu__26 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_26,), kwargs = {})
    %upsample_nearest2d : [num_users=1] = call_function[target=torch.ops.aten.upsample_nearest2d.vec](args = (%silu__26, None, [2.0, 2.0]), kwargs = {})
    %cat_5 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%upsample_nearest2d, %silu__19], 1), kwargs = {})
    %conv2d_27 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_5, %p_model_12_cv1_conv_weight, %p_model_12_cv1_conv_bias), kwargs = {})
    %silu__27 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_27,), kwargs = {})
    %split_with_sizes_4 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__27, [64, 64], 1), kwargs = {})
    %getitem_8 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_4, 0), kwargs = {})
    %getitem_9 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_4, 1), kwargs = {})
    %conv2d_28 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_9, %p_model_12_m_0_cv1_conv_weight, %p_model_12_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__28 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_28,), kwargs = {})
    %conv2d_29 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__28, %p_model_12_m_0_cv2_conv_weight, %p_model_12_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__29 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_29,), kwargs = {})
    %cat_6 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_8, %getitem_9, %silu__29], 1), kwargs = {})
    %conv2d_30 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_6, %p_model_12_cv2_conv_weight, %p_model_12_cv2_conv_bias), kwargs = {})
    %silu__30 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_30,), kwargs = {})
    %upsample_nearest2d_1 : [num_users=1] = call_function[target=torch.ops.aten.upsample_nearest2d.vec](args = (%silu__30, None, [2.0, 2.0]), kwargs = {})
    %cat_7 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%upsample_nearest2d_1, %silu__12], 1), kwargs = {})
    %conv2d_31 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_7, %p_model_15_cv1_conv_weight, %p_model_15_cv1_conv_bias), kwargs = {})
    %silu__31 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_31,), kwargs = {})
    %split_with_sizes_5 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__31, [32, 32], 1), kwargs = {})
    %getitem_10 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_5, 0), kwargs = {})
    %getitem_11 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_5, 1), kwargs = {})
    %conv2d_32 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_11, %p_model_15_m_0_cv1_conv_weight, %p_model_15_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__32 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_32,), kwargs = {})
    %conv2d_33 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__32, %p_model_15_m_0_cv2_conv_weight, %p_model_15_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__33 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_33,), kwargs = {})
    %cat_8 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_10, %getitem_11, %silu__33], 1), kwargs = {})
    %conv2d_34 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_8, %p_model_15_cv2_conv_weight, %p_model_15_cv2_conv_bias), kwargs = {})
    %silu__34 : [num_users=3] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_34,), kwargs = {})
    %conv2d_35 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_16_conv_weight, %p_model_16_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__35 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_35,), kwargs = {})
    %cat_9 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__35, %silu__30], 1), kwargs = {})
    %conv2d_36 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_9, %p_model_18_cv1_conv_weight, %p_model_18_cv1_conv_bias), kwargs = {})
    %silu__36 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_36,), kwargs = {})
    %split_with_sizes_6 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__36, [64, 64], 1), kwargs = {})
    %getitem_12 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_6, 0), kwargs = {})
    %getitem_13 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_6, 1), kwargs = {})
    %conv2d_37 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_13, %p_model_18_m_0_cv1_conv_weight, %p_model_18_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__37 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_37,), kwargs = {})
    %conv2d_38 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__37, %p_model_18_m_0_cv2_conv_weight, %p_model_18_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__38 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_38,), kwargs = {})
    %cat_10 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_12, %getitem_13, %silu__38], 1), kwargs = {})
    %conv2d_39 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_10, %p_model_18_cv2_conv_weight, %p_model_18_cv2_conv_bias), kwargs = {})
    %silu__39 : [num_users=3] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_39,), kwargs = {})
    %conv2d_40 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_19_conv_weight, %p_model_19_conv_bias, [2, 2], [1, 1]), kwargs = {})
    %silu__40 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_40,), kwargs = {})
    %cat_11 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%silu__40, %silu__26], 1), kwargs = {})
    %conv2d_41 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_11, %p_model_21_cv1_conv_weight, %p_model_21_cv1_conv_bias), kwargs = {})
    %silu__41 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_41,), kwargs = {})
    %split_with_sizes_7 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%silu__41, [128, 128], 1), kwargs = {})
    %getitem_14 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_7, 0), kwargs = {})
    %getitem_15 : [num_users=2] = call_function[target=operator.getitem](args = (%split_with_sizes_7, 1), kwargs = {})
    %conv2d_42 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%getitem_15, %p_model_21_m_0_cv1_conv_weight, %p_model_21_m_0_cv1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__42 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_42,), kwargs = {})
    %conv2d_43 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__42, %p_model_21_m_0_cv2_conv_weight, %p_model_21_m_0_cv2_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__43 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_43,), kwargs = {})
    %cat_12 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%getitem_14, %getitem_15, %silu__43], 1), kwargs = {})
    %conv2d_44 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%cat_12, %p_model_21_cv2_conv_weight, %p_model_21_cv2_conv_bias), kwargs = {})
    %silu__44 : [num_users=2] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_44,), kwargs = {})
    %conv2d_45 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_22_cv2_0_0_conv_weight, %p_model_22_cv2_0_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__45 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_45,), kwargs = {})
    %conv2d_46 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__45, %p_model_22_cv2_0_1_conv_weight, %p_model_22_cv2_0_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__46 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_46,), kwargs = {})
    %conv2d_47 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__46, %p_model_22_cv2_0_2_weight, %p_model_22_cv2_0_2_bias), kwargs = {})
    %conv2d_48 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__34, %p_model_22_cv3_0_0_conv_weight, %p_model_22_cv3_0_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__47 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_48,), kwargs = {})
    %conv2d_49 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__47, %p_model_22_cv3_0_1_conv_weight, %p_model_22_cv3_0_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__48 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_49,), kwargs = {})
    %conv2d_50 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__48, %p_model_22_cv3_0_2_weight, %p_model_22_cv3_0_2_bias), kwargs = {})
    %cat_13 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_47, %conv2d_50], 1), kwargs = {})
    %conv2d_51 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_22_cv2_1_0_conv_weight, %p_model_22_cv2_1_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__49 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_51,), kwargs = {})
    %conv2d_52 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__49, %p_model_22_cv2_1_1_conv_weight, %p_model_22_cv2_1_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__50 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_52,), kwargs = {})
    %conv2d_53 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__50, %p_model_22_cv2_1_2_weight, %p_model_22_cv2_1_2_bias), kwargs = {})
    %conv2d_54 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__39, %p_model_22_cv3_1_0_conv_weight, %p_model_22_cv3_1_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__51 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_54,), kwargs = {})
    %conv2d_55 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__51, %p_model_22_cv3_1_1_conv_weight, %p_model_22_cv3_1_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__52 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_55,), kwargs = {})
    %conv2d_56 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__52, %p_model_22_cv3_1_2_weight, %p_model_22_cv3_1_2_bias), kwargs = {})
    %cat_14 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_53, %conv2d_56], 1), kwargs = {})
    %conv2d_57 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__44, %p_model_22_cv2_2_0_conv_weight, %p_model_22_cv2_2_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__53 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_57,), kwargs = {})
    %conv2d_58 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__53, %p_model_22_cv2_2_1_conv_weight, %p_model_22_cv2_2_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__54 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_58,), kwargs = {})
    %conv2d_59 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__54, %p_model_22_cv2_2_2_weight, %p_model_22_cv2_2_2_bias), kwargs = {})
    %conv2d_60 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__44, %p_model_22_cv3_2_0_conv_weight, %p_model_22_cv3_2_0_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__55 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_60,), kwargs = {})
    %conv2d_61 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__55, %p_model_22_cv3_2_1_conv_weight, %p_model_22_cv3_2_1_conv_bias, [1, 1], [1, 1]), kwargs = {})
    %silu__56 : [num_users=1] = call_function[target=torch.ops.aten.silu_.default](args = (%conv2d_61,), kwargs = {})
    %conv2d_62 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%silu__56, %p_model_22_cv3_2_2_weight, %p_model_22_cv3_2_2_bias), kwargs = {})
    %cat_15 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%conv2d_59, %conv2d_62], 1), kwargs = {})
    %view : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_13, [1, 144, -1]), kwargs = {})
    %view_1 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_14, [1, 144, -1]), kwargs = {})
    %view_2 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%cat_15, [1, 144, -1]), kwargs = {})
    %cat_16 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%view, %view_1, %view_2], 2), kwargs = {})
    %split_with_sizes_8 : [num_users=2] = call_function[target=torch.ops.aten.split_with_sizes.default](args = (%cat_16, [64, 80], 1), kwargs = {})
    %getitem_16 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_8, 0), kwargs = {})
    %getitem_17 : [num_users=1] = call_function[target=operator.getitem](args = (%split_with_sizes_8, 1), kwargs = {})
    %view_3 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%getitem_16, [1, 4, 16, 8400]), kwargs = {})
    %transpose : [num_users=1] = call_function[target=torch.ops.aten.transpose.int](args = (%view_3, 2, 1), kwargs = {})
    %softmax : [num_users=1] = call_function[target=torch.ops.aten.softmax.int](args = (%transpose, 1), kwargs = {})
    %conv2d_63 : [num_users=1] = call_function[target=torch.ops.aten.conv2d.default](args = (%softmax, %p_model_22_dfl_conv_weight), kwargs = {})
    %view_4 : [num_users=1] = call_function[target=torch.ops.aten.view.default](args = (%conv2d_63, [1, 4, 8400]), kwargs = {})
    %chunk : [num_users=2] = call_function[target=torch.ops.aten.chunk.default](args = (%view_4, 2, 1), kwargs = {})
    %getitem_18 : [num_users=1] = call_function[target=operator.getitem](args = (%chunk, 0), kwargs = {})
    %getitem_19 : [num_users=1] = call_function[target=operator.getitem](args = (%chunk, 1), kwargs = {})
    %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_prop_tensor_constant2, %getitem_18), kwargs = {})
    %add_6 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%_prop_tensor_constant2, %getitem_19), kwargs = {})
    %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub, %add_6), kwargs = {})
    %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_7, 2), kwargs = {})
    %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_6, %sub), kwargs = {})
    %cat_17 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%div, %sub_1], 1), kwargs = {})
    %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%cat_17, %c_model_22_strides), kwargs = {})
    %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%getitem_17,), kwargs = {})
    %cat_18 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%mul, %sigmoid], 1), kwargs = {})
    return (cat_18,)

---


ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_weight'), target='model.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_bias'), target='model.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_weight'), target='model.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_bias'), target='model.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_weight'), target='model.2.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_bias'), target='model.2.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_weight'), target='model.2.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_bias'), target='model.2.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_weight'), target='model.2.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_bias'), target='model.2.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_weight'), target='model.2.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_bias'), target='model.2.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_weight'), target='model.3.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_bias'), target='model.3.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_weight'), target='model.4.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_bias'), target='model.4.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_weight'), target='model.4.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_bias'), target='model.4.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_weight'), target='model.4.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_bias'), target='model.4.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_weight'), target='model.4.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_bias'), target='model.4.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_weight'), target='model.4.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_bias'), target='model.4.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_weight'), target='model.4.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_bias'), target='model.4.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_weight'), target='model.5.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_bias'), target='model.5.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_weight'), target='model.6.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_bias'), target='model.6.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_weight'), target='model.6.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_bias'), target='model.6.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_weight'), target='model.6.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_bias'), target='model.6.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_weight'), target='model.6.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_bias'), target='model.6.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_weight'), target='model.6.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_bias'), target='model.6.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_weight'), target='model.6.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_bias'), target='model.6.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_weight'), target='model.7.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_bias'), target='model.7.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_weight'), target='model.8.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_bias'), target='model.8.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_weight'), target='model.8.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_bias'), target='model.8.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_weight'), target='model.8.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_bias'), target='model.8.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_weight'), target='model.8.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_bias'), target='model.8.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_weight'), target='model.9.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_bias'), target='model.9.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_weight'), target='model.9.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_bias'), target='model.9.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_weight'), target='model.12.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_bias'), target='model.12.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_weight'), target='model.12.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_bias'), target='model.12.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_weight'), target='model.12.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_bias'), target='model.12.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_weight'), target='model.12.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_bias'), target='model.12.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_weight'), target='model.15.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_bias'), target='model.15.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_weight'), target='model.15.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_bias'), target='model.15.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_weight'), target='model.15.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_bias'), target='model.15.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_weight'), target='model.15.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_bias'), target='model.15.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_weight'), target='model.16.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_bias'), target='model.16.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_weight'), target='model.18.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_bias'), target='model.18.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_weight'), target='model.18.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_bias'), target='model.18.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_weight'), target='model.18.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_bias'), target='model.18.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_weight'), target='model.18.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_bias'), target='model.18.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_weight'), target='model.19.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_bias'), target='model.19.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_weight'), target='model.21.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_bias'), target='model.21.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_weight'), target='model.21.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_bias'), target='model.21.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_weight'), target='model.21.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_bias'), target='model.21.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_weight'), target='model.21.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_bias'), target='model.21.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_weight'), target='model.22.cv2.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_bias'), target='model.22.cv2.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_weight'), target='model.22.cv2.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_bias'), target='model.22.cv2.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_weight'), target='model.22.cv2.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_bias'), target='model.22.cv2.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_weight'), target='model.22.cv2.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_bias'), target='model.22.cv2.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_weight'), target='model.22.cv2.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_bias'), target='model.22.cv2.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_weight'), target='model.22.cv2.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_bias'), target='model.22.cv2.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_weight'), target='model.22.cv2.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_bias'), target='model.22.cv2.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_weight'), target='model.22.cv2.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_bias'), target='model.22.cv2.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_weight'), target='model.22.cv2.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_bias'), target='model.22.cv2.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_weight'), target='model.22.cv3.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_bias'), target='model.22.cv3.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_weight'), target='model.22.cv3.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_bias'), target='model.22.cv3.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_weight'), target='model.22.cv3.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_bias'), target='model.22.cv3.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_weight'), target='model.22.cv3.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_bias'), target='model.22.cv3.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_weight'), target='model.22.cv3.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_bias'), target='model.22.cv3.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_weight'), target='model.22.cv3.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_bias'), target='model.22.cv3.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_weight'), target='model.22.cv3.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_bias'), target='model.22.cv3.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_weight'), target='model.22.cv3.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_bias'), target='model.22.cv3.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_weight'), target='model.22.cv3.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_bias'), target='model.22.cv3.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_dfl_conv_weight'), target='model.22.dfl.conv.weight', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_strides'), target='model.22.strides', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='_prop_tensor_constant2'), target='_prop_tensor_constant2', persistent=True), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='x'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='cat_18'), target=None)])

And edge dialect:

print(edge_dialect_program.exported_program().graph_module.graph)
print(edge_dialect_program.exported_program().graph_signature)

Output:

graph():
    %p_model_0_conv_weight : [num_users=1] = placeholder[target=p_model_0_conv_weight]
    %p_model_0_conv_bias : [num_users=1] = placeholder[target=p_model_0_conv_bias]
    %p_model_1_conv_weight : [num_users=1] = placeholder[target=p_model_1_conv_weight]
    %p_model_1_conv_bias : [num_users=1] = placeholder[target=p_model_1_conv_bias]
    %p_model_2_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv1_conv_weight]
    %p_model_2_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv1_conv_bias]
    %p_model_2_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_cv2_conv_weight]
    %p_model_2_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_cv2_conv_bias]
    %p_model_2_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_weight]
    %p_model_2_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv1_conv_bias]
    %p_model_2_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_weight]
    %p_model_2_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_2_m_0_cv2_conv_bias]
    %p_model_3_conv_weight : [num_users=1] = placeholder[target=p_model_3_conv_weight]
    %p_model_3_conv_bias : [num_users=1] = placeholder[target=p_model_3_conv_bias]
    %p_model_4_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv1_conv_weight]
    %p_model_4_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv1_conv_bias]
    %p_model_4_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_cv2_conv_weight]
    %p_model_4_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_cv2_conv_bias]
    %p_model_4_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_weight]
    %p_model_4_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv1_conv_bias]
    %p_model_4_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_weight]
    %p_model_4_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_0_cv2_conv_bias]
    %p_model_4_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_weight]
    %p_model_4_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv1_conv_bias]
    %p_model_4_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_weight]
    %p_model_4_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_4_m_1_cv2_conv_bias]
    %p_model_5_conv_weight : [num_users=1] = placeholder[target=p_model_5_conv_weight]
    %p_model_5_conv_bias : [num_users=1] = placeholder[target=p_model_5_conv_bias]
    %p_model_6_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv1_conv_weight]
    %p_model_6_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv1_conv_bias]
    %p_model_6_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_cv2_conv_weight]
    %p_model_6_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_cv2_conv_bias]
    %p_model_6_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_weight]
    %p_model_6_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv1_conv_bias]
    %p_model_6_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_weight]
    %p_model_6_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_0_cv2_conv_bias]
    %p_model_6_m_1_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_weight]
    %p_model_6_m_1_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv1_conv_bias]
    %p_model_6_m_1_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_weight]
    %p_model_6_m_1_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_6_m_1_cv2_conv_bias]
    %p_model_7_conv_weight : [num_users=1] = placeholder[target=p_model_7_conv_weight]
    %p_model_7_conv_bias : [num_users=1] = placeholder[target=p_model_7_conv_bias]
    %p_model_8_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv1_conv_weight]
    %p_model_8_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv1_conv_bias]
    %p_model_8_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_cv2_conv_weight]
    %p_model_8_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_cv2_conv_bias]
    %p_model_8_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_weight]
    %p_model_8_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv1_conv_bias]
    %p_model_8_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_weight]
    %p_model_8_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_8_m_0_cv2_conv_bias]
    %p_model_9_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv1_conv_weight]
    %p_model_9_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv1_conv_bias]
    %p_model_9_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_9_cv2_conv_weight]
    %p_model_9_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_9_cv2_conv_bias]
    %p_model_12_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv1_conv_weight]
    %p_model_12_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv1_conv_bias]
    %p_model_12_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_cv2_conv_weight]
    %p_model_12_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_cv2_conv_bias]
    %p_model_12_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_weight]
    %p_model_12_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv1_conv_bias]
    %p_model_12_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_weight]
    %p_model_12_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_12_m_0_cv2_conv_bias]
    %p_model_15_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv1_conv_weight]
    %p_model_15_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv1_conv_bias]
    %p_model_15_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_cv2_conv_weight]
    %p_model_15_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_cv2_conv_bias]
    %p_model_15_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_weight]
    %p_model_15_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv1_conv_bias]
    %p_model_15_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_weight]
    %p_model_15_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_15_m_0_cv2_conv_bias]
    %p_model_16_conv_weight : [num_users=1] = placeholder[target=p_model_16_conv_weight]
    %p_model_16_conv_bias : [num_users=1] = placeholder[target=p_model_16_conv_bias]
    %p_model_18_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv1_conv_weight]
    %p_model_18_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv1_conv_bias]
    %p_model_18_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_cv2_conv_weight]
    %p_model_18_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_cv2_conv_bias]
    %p_model_18_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_weight]
    %p_model_18_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv1_conv_bias]
    %p_model_18_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_weight]
    %p_model_18_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_18_m_0_cv2_conv_bias]
    %p_model_19_conv_weight : [num_users=1] = placeholder[target=p_model_19_conv_weight]
    %p_model_19_conv_bias : [num_users=1] = placeholder[target=p_model_19_conv_bias]
    %p_model_21_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv1_conv_weight]
    %p_model_21_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv1_conv_bias]
    %p_model_21_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_cv2_conv_weight]
    %p_model_21_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_cv2_conv_bias]
    %p_model_21_m_0_cv1_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_weight]
    %p_model_21_m_0_cv1_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv1_conv_bias]
    %p_model_21_m_0_cv2_conv_weight : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_weight]
    %p_model_21_m_0_cv2_conv_bias : [num_users=1] = placeholder[target=p_model_21_m_0_cv2_conv_bias]
    %p_model_22_cv2_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_weight]
    %p_model_22_cv2_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_0_conv_bias]
    %p_model_22_cv2_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_weight]
    %p_model_22_cv2_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_1_conv_bias]
    %p_model_22_cv2_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_weight]
    %p_model_22_cv2_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_0_2_bias]
    %p_model_22_cv2_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_weight]
    %p_model_22_cv2_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_0_conv_bias]
    %p_model_22_cv2_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_weight]
    %p_model_22_cv2_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_1_conv_bias]
    %p_model_22_cv2_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_weight]
    %p_model_22_cv2_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_1_2_bias]
    %p_model_22_cv2_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_weight]
    %p_model_22_cv2_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_0_conv_bias]
    %p_model_22_cv2_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_weight]
    %p_model_22_cv2_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_1_conv_bias]
    %p_model_22_cv2_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_weight]
    %p_model_22_cv2_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv2_2_2_bias]
    %p_model_22_cv3_0_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_weight]
    %p_model_22_cv3_0_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_0_conv_bias]
    %p_model_22_cv3_0_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_weight]
    %p_model_22_cv3_0_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_1_conv_bias]
    %p_model_22_cv3_0_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_weight]
    %p_model_22_cv3_0_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_0_2_bias]
    %p_model_22_cv3_1_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_weight]
    %p_model_22_cv3_1_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_0_conv_bias]
    %p_model_22_cv3_1_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_weight]
    %p_model_22_cv3_1_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_1_conv_bias]
    %p_model_22_cv3_1_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_weight]
    %p_model_22_cv3_1_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_1_2_bias]
    %p_model_22_cv3_2_0_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_weight]
    %p_model_22_cv3_2_0_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_0_conv_bias]
    %p_model_22_cv3_2_1_conv_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_weight]
    %p_model_22_cv3_2_1_conv_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_1_conv_bias]
    %p_model_22_cv3_2_2_weight : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_weight]
    %p_model_22_cv3_2_2_bias : [num_users=1] = placeholder[target=p_model_22_cv3_2_2_bias]
    %p_model_22_dfl_conv_weight : [num_users=1] = placeholder[target=p_model_22_dfl_conv_weight]
    %c__prop_tensor_constant2 : [num_users=2] = placeholder[target=c__prop_tensor_constant2]
    %c_model_22_strides : [num_users=1] = placeholder[target=c_model_22_strides]
    %_lifted_tensor_constant0 : [num_users=1] = placeholder[target=_lifted_tensor_constant0]
    %_lifted_tensor_constant1 : [num_users=1] = placeholder[target=_lifted_tensor_constant1]
    %_lifted_tensor_constant2 : [num_users=1] = placeholder[target=_lifted_tensor_constant2]
    %_lifted_tensor_constant3 : [num_users=1] = placeholder[target=_lifted_tensor_constant3]
    %_lifted_tensor_constant4 : [num_users=1] = placeholder[target=_lifted_tensor_constant4]
    %_lifted_tensor_constant5 : [num_users=1] = placeholder[target=_lifted_tensor_constant5]
    %_lifted_tensor_constant6 : [num_users=1] = placeholder[target=_lifted_tensor_constant6]
    %_lifted_tensor_constant7 : [num_users=1] = placeholder[target=_lifted_tensor_constant7]
    %_lifted_tensor_constant8 : [num_users=1] = placeholder[target=_lifted_tensor_constant8]
    %x : [num_users=1] = placeholder[target=x]
    %aten_convolution_default : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%x, %p_model_0_conv_weight, %p_model_0_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default,), kwargs = {})
    %aten_mul_tensor : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default, %aten_sigmoid_default), kwargs = {})
    %aten_convolution_default_1 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor, %p_model_1_conv_weight, %p_model_1_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_1,), kwargs = {})
    %aten_mul_tensor_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_1, %aten_sigmoid_default_1), kwargs = {})
    %aten_convolution_default_2 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_1, %p_model_2_cv1_conv_weight, %p_model_2_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_2,), kwargs = {})
    %aten_mul_tensor_2 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_2, %aten_sigmoid_default_2), kwargs = {})
    %aten_split_with_sizes_copy_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_2, [16, 16], 1), kwargs = {})
    %getitem : [num_users=3] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default, 1), kwargs = {})
    %aten_convolution_default_3 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem, %p_model_2_m_0_cv1_conv_weight, %p_model_2_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_3,), kwargs = {})
    %aten_mul_tensor_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_3, %aten_sigmoid_default_3), kwargs = {})
    %aten_convolution_default_4 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_3, %p_model_2_m_0_cv2_conv_weight, %p_model_2_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_4,), kwargs = {})
    %aten_mul_tensor_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_4, %aten_sigmoid_default_4), kwargs = {})
    %aten_add_tensor : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%getitem, %aten_mul_tensor_4), kwargs = {})
    %aten_split_with_sizes_copy_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_2, [16, 16], 1), kwargs = {})
    %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_1, 0), kwargs = {})
    %aten_cat_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_1, %getitem, %aten_add_tensor], 1), kwargs = {})
    %aten_convolution_default_5 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default, %p_model_2_cv2_conv_weight, %p_model_2_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_5 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_5,), kwargs = {})
    %aten_mul_tensor_5 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_5, %aten_sigmoid_default_5), kwargs = {})
    %aten_convolution_default_6 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_5, %p_model_3_conv_weight, %p_model_3_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_6 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_6,), kwargs = {})
    %aten_mul_tensor_6 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_6, %aten_sigmoid_default_6), kwargs = {})
    %aten_convolution_default_7 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_6, %p_model_4_cv1_conv_weight, %p_model_4_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_7 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_7,), kwargs = {})
    %aten_mul_tensor_7 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_7, %aten_sigmoid_default_7), kwargs = {})
    %aten_split_with_sizes_copy_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_7, [32, 32], 1), kwargs = {})
    %getitem_2 : [num_users=3] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_2, 1), kwargs = {})
    %aten_convolution_default_8 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_2, %p_model_4_m_0_cv1_conv_weight, %p_model_4_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_8 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_8,), kwargs = {})
    %aten_mul_tensor_8 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_8, %aten_sigmoid_default_8), kwargs = {})
    %aten_convolution_default_9 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_8, %p_model_4_m_0_cv2_conv_weight, %p_model_4_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_9 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_9,), kwargs = {})
    %aten_mul_tensor_9 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_9, %aten_sigmoid_default_9), kwargs = {})
    %aten_add_tensor_1 : [num_users=3] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%getitem_2, %aten_mul_tensor_9), kwargs = {})
    %aten_convolution_default_10 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_add_tensor_1, %p_model_4_m_1_cv1_conv_weight, %p_model_4_m_1_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_10 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_10,), kwargs = {})
    %aten_mul_tensor_10 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_10, %aten_sigmoid_default_10), kwargs = {})
    %aten_convolution_default_11 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_10, %p_model_4_m_1_cv2_conv_weight, %p_model_4_m_1_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_11 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_11,), kwargs = {})
    %aten_mul_tensor_11 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_11, %aten_sigmoid_default_11), kwargs = {})
    %aten_add_tensor_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%aten_add_tensor_1, %aten_mul_tensor_11), kwargs = {})
    %aten_split_with_sizes_copy_default_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_7, [32, 32], 1), kwargs = {})
    %getitem_3 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_3, 0), kwargs = {})
    %aten_cat_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_3, %getitem_2, %aten_add_tensor_1, %aten_add_tensor_2], 1), kwargs = {})
    %aten_convolution_default_12 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_1, %p_model_4_cv2_conv_weight, %p_model_4_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_12 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_12,), kwargs = {})
    %aten_mul_tensor_12 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_12, %aten_sigmoid_default_12), kwargs = {})
    %aten_convolution_default_13 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_12, %p_model_5_conv_weight, %p_model_5_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_13 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_13,), kwargs = {})
    %aten_mul_tensor_13 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_13, %aten_sigmoid_default_13), kwargs = {})
    %aten_convolution_default_14 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_13, %p_model_6_cv1_conv_weight, %p_model_6_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_14 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_14,), kwargs = {})
    %aten_mul_tensor_14 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_14, %aten_sigmoid_default_14), kwargs = {})
    %aten_split_with_sizes_copy_default_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_14, [64, 64], 1), kwargs = {})
    %getitem_4 : [num_users=3] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_4, 1), kwargs = {})
    %aten_convolution_default_15 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_4, %p_model_6_m_0_cv1_conv_weight, %p_model_6_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_15 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_15,), kwargs = {})
    %aten_mul_tensor_15 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_15, %aten_sigmoid_default_15), kwargs = {})
    %aten_convolution_default_16 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_15, %p_model_6_m_0_cv2_conv_weight, %p_model_6_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_16 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_16,), kwargs = {})
    %aten_mul_tensor_16 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_16, %aten_sigmoid_default_16), kwargs = {})
    %aten_add_tensor_3 : [num_users=3] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%getitem_4, %aten_mul_tensor_16), kwargs = {})
    %aten_convolution_default_17 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_add_tensor_3, %p_model_6_m_1_cv1_conv_weight, %p_model_6_m_1_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_17 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_17,), kwargs = {})
    %aten_mul_tensor_17 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_17, %aten_sigmoid_default_17), kwargs = {})
    %aten_convolution_default_18 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_17, %p_model_6_m_1_cv2_conv_weight, %p_model_6_m_1_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_18 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_18,), kwargs = {})
    %aten_mul_tensor_18 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_18, %aten_sigmoid_default_18), kwargs = {})
    %aten_add_tensor_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%aten_add_tensor_3, %aten_mul_tensor_18), kwargs = {})
    %aten_split_with_sizes_copy_default_5 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_14, [64, 64], 1), kwargs = {})
    %getitem_5 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_5, 0), kwargs = {})
    %aten_cat_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_5, %getitem_4, %aten_add_tensor_3, %aten_add_tensor_4], 1), kwargs = {})
    %aten_convolution_default_19 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_2, %p_model_6_cv2_conv_weight, %p_model_6_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_19 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_19,), kwargs = {})
    %aten_mul_tensor_19 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_19, %aten_sigmoid_default_19), kwargs = {})
    %aten_convolution_default_20 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_19, %p_model_7_conv_weight, %p_model_7_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_20 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_20,), kwargs = {})
    %aten_mul_tensor_20 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_20, %aten_sigmoid_default_20), kwargs = {})
    %aten_convolution_default_21 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_20, %p_model_8_cv1_conv_weight, %p_model_8_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_21 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_21,), kwargs = {})
    %aten_mul_tensor_21 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_21, %aten_sigmoid_default_21), kwargs = {})
    %aten_split_with_sizes_copy_default_6 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_21, [128, 128], 1), kwargs = {})
    %getitem_6 : [num_users=3] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_6, 1), kwargs = {})
    %aten_convolution_default_22 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_6, %p_model_8_m_0_cv1_conv_weight, %p_model_8_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_22 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_22,), kwargs = {})
    %aten_mul_tensor_22 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_22, %aten_sigmoid_default_22), kwargs = {})
    %aten_convolution_default_23 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_22, %p_model_8_m_0_cv2_conv_weight, %p_model_8_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_23 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_23,), kwargs = {})
    %aten_mul_tensor_23 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_23, %aten_sigmoid_default_23), kwargs = {})
    %aten_add_tensor_5 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%getitem_6, %aten_mul_tensor_23), kwargs = {})
    %aten_split_with_sizes_copy_default_7 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_21, [128, 128], 1), kwargs = {})
    %getitem_7 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_7, 0), kwargs = {})
    %aten_cat_default_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_7, %getitem_6, %aten_add_tensor_5], 1), kwargs = {})
    %aten_convolution_default_24 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_3, %p_model_8_cv2_conv_weight, %p_model_8_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_24 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_24,), kwargs = {})
    %aten_mul_tensor_24 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_24, %aten_sigmoid_default_24), kwargs = {})
    %aten_convolution_default_25 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_24, %p_model_9_cv1_conv_weight, %p_model_9_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_25 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_25,), kwargs = {})
    %aten_mul_tensor_25 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_25, %aten_sigmoid_default_25), kwargs = {})
    %aten_max_pool2d_with_indices_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.max_pool2d_with_indices.default](args = (%aten_mul_tensor_25, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %getitem_8 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_max_pool2d_with_indices_default, 0), kwargs = {})
    %aten_max_pool2d_with_indices_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.max_pool2d_with_indices.default](args = (%getitem_8, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %getitem_9 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_max_pool2d_with_indices_default_1, 0), kwargs = {})
    %aten_max_pool2d_with_indices_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.max_pool2d_with_indices.default](args = (%getitem_9, [5, 5], [1, 1], [2, 2]), kwargs = {})
    %getitem_10 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_max_pool2d_with_indices_default_2, 0), kwargs = {})
    %aten_cat_default_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_mul_tensor_25, %getitem_8, %getitem_9, %getitem_10], 1), kwargs = {})
    %aten_convolution_default_26 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_4, %p_model_9_cv2_conv_weight, %p_model_9_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_26 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_26,), kwargs = {})
    %aten_mul_tensor_26 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_26, %aten_sigmoid_default_26), kwargs = {})
    %_tensor_constant0 : [num_users=1] = get_attr[target=_tensor_constant0]
    %aten_add_tensor_6 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%_tensor_constant0, %_lifted_tensor_constant0), kwargs = {})
    %aten_mul_tensor_27 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_add_tensor_6, %_lifted_tensor_constant1), kwargs = {})
    %dim_order_ops__to_dim_order_copy_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.dim_order_ops._to_dim_order_copy.default](args = (%aten_mul_tensor_27,), kwargs = {dtype: torch.int64, dim_order: [0]})
    %aten_unsqueeze_copy_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.unsqueeze_copy.default](args = (%dim_order_ops__to_dim_order_copy_default, -1), kwargs = {})
    %_tensor_constant1 : [num_users=1] = get_attr[target=_tensor_constant1]
    %aten_add_tensor_7 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%_tensor_constant1, %_lifted_tensor_constant2), kwargs = {})
    %aten_mul_tensor_28 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_add_tensor_7, %_lifted_tensor_constant3), kwargs = {})
    %dim_order_ops__to_dim_order_copy_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.dim_order_ops._to_dim_order_copy.default](args = (%aten_mul_tensor_28,), kwargs = {dtype: torch.int64, dim_order: [0]})
    %aten_index_tensor : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.index.Tensor](args = (%aten_mul_tensor_26, [None, None, %aten_unsqueeze_copy_default, %dim_order_ops__to_dim_order_copy_default_1]), kwargs = {})
    %aten_cat_default_5 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_index_tensor, %aten_mul_tensor_19], 1), kwargs = {})
    %aten_convolution_default_27 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_5, %p_model_12_cv1_conv_weight, %p_model_12_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_27 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_27,), kwargs = {})
    %aten_mul_tensor_29 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_27, %aten_sigmoid_default_27), kwargs = {})
    %aten_split_with_sizes_copy_default_8 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_29, [64, 64], 1), kwargs = {})
    %getitem_11 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_8, 1), kwargs = {})
    %aten_convolution_default_28 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_11, %p_model_12_m_0_cv1_conv_weight, %p_model_12_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_28 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_28,), kwargs = {})
    %aten_mul_tensor_30 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_28, %aten_sigmoid_default_28), kwargs = {})
    %aten_convolution_default_29 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_30, %p_model_12_m_0_cv2_conv_weight, %p_model_12_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_29 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_29,), kwargs = {})
    %aten_mul_tensor_31 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_29, %aten_sigmoid_default_29), kwargs = {})
    %aten_split_with_sizes_copy_default_9 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_29, [64, 64], 1), kwargs = {})
    %getitem_12 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_9, 0), kwargs = {})
    %aten_cat_default_6 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_12, %getitem_11, %aten_mul_tensor_31], 1), kwargs = {})
    %aten_convolution_default_30 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_6, %p_model_12_cv2_conv_weight, %p_model_12_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_30 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_30,), kwargs = {})
    %aten_mul_tensor_32 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_30, %aten_sigmoid_default_30), kwargs = {})
    %_tensor_constant2 : [num_users=1] = get_attr[target=_tensor_constant2]
    %aten_add_tensor_8 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%_tensor_constant2, %_lifted_tensor_constant4), kwargs = {})
    %aten_mul_tensor_33 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_add_tensor_8, %_lifted_tensor_constant5), kwargs = {})
    %dim_order_ops__to_dim_order_copy_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.dim_order_ops._to_dim_order_copy.default](args = (%aten_mul_tensor_33,), kwargs = {dtype: torch.int64, dim_order: [0]})
    %aten_unsqueeze_copy_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.unsqueeze_copy.default](args = (%dim_order_ops__to_dim_order_copy_default_2, -1), kwargs = {})
    %_tensor_constant3 : [num_users=1] = get_attr[target=_tensor_constant3]
    %aten_add_tensor_9 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%_tensor_constant3, %_lifted_tensor_constant6), kwargs = {})
    %aten_mul_tensor_34 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_add_tensor_9, %_lifted_tensor_constant7), kwargs = {})
    %dim_order_ops__to_dim_order_copy_default_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.dim_order_ops._to_dim_order_copy.default](args = (%aten_mul_tensor_34,), kwargs = {dtype: torch.int64, dim_order: [0]})
    %aten_index_tensor_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.index.Tensor](args = (%aten_mul_tensor_32, [None, None, %aten_unsqueeze_copy_default_1, %dim_order_ops__to_dim_order_copy_default_3]), kwargs = {})
    %aten_cat_default_7 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_index_tensor_1, %aten_mul_tensor_12], 1), kwargs = {})
    %aten_convolution_default_31 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_7, %p_model_15_cv1_conv_weight, %p_model_15_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_31 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_31,), kwargs = {})
    %aten_mul_tensor_35 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_31, %aten_sigmoid_default_31), kwargs = {})
    %aten_split_with_sizes_copy_default_10 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_35, [32, 32], 1), kwargs = {})
    %getitem_13 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_10, 1), kwargs = {})
    %aten_convolution_default_32 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_13, %p_model_15_m_0_cv1_conv_weight, %p_model_15_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_32 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_32,), kwargs = {})
    %aten_mul_tensor_36 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_32, %aten_sigmoid_default_32), kwargs = {})
    %aten_convolution_default_33 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_36, %p_model_15_m_0_cv2_conv_weight, %p_model_15_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_33 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_33,), kwargs = {})
    %aten_mul_tensor_37 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_33, %aten_sigmoid_default_33), kwargs = {})
    %aten_split_with_sizes_copy_default_11 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_35, [32, 32], 1), kwargs = {})
    %getitem_14 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_11, 0), kwargs = {})
    %aten_cat_default_8 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_14, %getitem_13, %aten_mul_tensor_37], 1), kwargs = {})
    %aten_convolution_default_34 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_8, %p_model_15_cv2_conv_weight, %p_model_15_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_34 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_34,), kwargs = {})
    %aten_mul_tensor_38 : [num_users=3] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_34, %aten_sigmoid_default_34), kwargs = {})
    %aten_convolution_default_35 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_38, %p_model_16_conv_weight, %p_model_16_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_35 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_35,), kwargs = {})
    %aten_mul_tensor_39 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_35, %aten_sigmoid_default_35), kwargs = {})
    %aten_cat_default_9 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_mul_tensor_39, %aten_mul_tensor_32], 1), kwargs = {})
    %aten_convolution_default_36 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_9, %p_model_18_cv1_conv_weight, %p_model_18_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_36 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_36,), kwargs = {})
    %aten_mul_tensor_40 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_36, %aten_sigmoid_default_36), kwargs = {})
    %aten_split_with_sizes_copy_default_12 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_40, [64, 64], 1), kwargs = {})
    %getitem_15 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_12, 1), kwargs = {})
    %aten_convolution_default_37 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_15, %p_model_18_m_0_cv1_conv_weight, %p_model_18_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_37 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_37,), kwargs = {})
    %aten_mul_tensor_41 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_37, %aten_sigmoid_default_37), kwargs = {})
    %aten_convolution_default_38 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_41, %p_model_18_m_0_cv2_conv_weight, %p_model_18_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_38 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_38,), kwargs = {})
    %aten_mul_tensor_42 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_38, %aten_sigmoid_default_38), kwargs = {})
    %aten_split_with_sizes_copy_default_13 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_40, [64, 64], 1), kwargs = {})
    %getitem_16 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_13, 0), kwargs = {})
    %aten_cat_default_10 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_16, %getitem_15, %aten_mul_tensor_42], 1), kwargs = {})
    %aten_convolution_default_39 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_10, %p_model_18_cv2_conv_weight, %p_model_18_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_39 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_39,), kwargs = {})
    %aten_mul_tensor_43 : [num_users=3] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_39, %aten_sigmoid_default_39), kwargs = {})
    %aten_convolution_default_40 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_43, %p_model_19_conv_weight, %p_model_19_conv_bias, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_40 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_40,), kwargs = {})
    %aten_mul_tensor_44 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_40, %aten_sigmoid_default_40), kwargs = {})
    %aten_cat_default_11 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_mul_tensor_44, %aten_mul_tensor_26], 1), kwargs = {})
    %aten_convolution_default_41 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_11, %p_model_21_cv1_conv_weight, %p_model_21_cv1_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_41 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_41,), kwargs = {})
    %aten_mul_tensor_45 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_41, %aten_sigmoid_default_41), kwargs = {})
    %aten_split_with_sizes_copy_default_14 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_45, [128, 128], 1), kwargs = {})
    %getitem_17 : [num_users=2] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_14, 1), kwargs = {})
    %aten_convolution_default_42 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%getitem_17, %p_model_21_m_0_cv1_conv_weight, %p_model_21_m_0_cv1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_42 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_42,), kwargs = {})
    %aten_mul_tensor_46 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_42, %aten_sigmoid_default_42), kwargs = {})
    %aten_convolution_default_43 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_46, %p_model_21_m_0_cv2_conv_weight, %p_model_21_m_0_cv2_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_43 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_43,), kwargs = {})
    %aten_mul_tensor_47 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_43, %aten_sigmoid_default_43), kwargs = {})
    %aten_split_with_sizes_copy_default_15 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_mul_tensor_45, [128, 128], 1), kwargs = {})
    %getitem_18 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_15, 0), kwargs = {})
    %aten_cat_default_12 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%getitem_18, %getitem_17, %aten_mul_tensor_47], 1), kwargs = {})
    %aten_convolution_default_44 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_cat_default_12, %p_model_21_cv2_conv_weight, %p_model_21_cv2_conv_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_44 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_44,), kwargs = {})
    %aten_mul_tensor_48 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_44, %aten_sigmoid_default_44), kwargs = {})
    %aten_convolution_default_45 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_38, %p_model_22_cv2_0_0_conv_weight, %p_model_22_cv2_0_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_45 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_45,), kwargs = {})
    %aten_mul_tensor_49 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_45, %aten_sigmoid_default_45), kwargs = {})
    %aten_convolution_default_46 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_49, %p_model_22_cv2_0_1_conv_weight, %p_model_22_cv2_0_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_46 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_46,), kwargs = {})
    %aten_mul_tensor_50 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_46, %aten_sigmoid_default_46), kwargs = {})
    %aten_convolution_default_47 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_50, %p_model_22_cv2_0_2_weight, %p_model_22_cv2_0_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_convolution_default_48 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_38, %p_model_22_cv3_0_0_conv_weight, %p_model_22_cv3_0_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_47 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_48,), kwargs = {})
    %aten_mul_tensor_51 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_48, %aten_sigmoid_default_47), kwargs = {})
    %aten_convolution_default_49 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_51, %p_model_22_cv3_0_1_conv_weight, %p_model_22_cv3_0_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_48 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_49,), kwargs = {})
    %aten_mul_tensor_52 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_49, %aten_sigmoid_default_48), kwargs = {})
    %aten_convolution_default_50 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_52, %p_model_22_cv3_0_2_weight, %p_model_22_cv3_0_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_cat_default_13 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_convolution_default_47, %aten_convolution_default_50], 1), kwargs = {})
    %aten_convolution_default_51 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_43, %p_model_22_cv2_1_0_conv_weight, %p_model_22_cv2_1_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_49 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_51,), kwargs = {})
    %aten_mul_tensor_53 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_51, %aten_sigmoid_default_49), kwargs = {})
    %aten_convolution_default_52 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_53, %p_model_22_cv2_1_1_conv_weight, %p_model_22_cv2_1_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_50 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_52,), kwargs = {})
    %aten_mul_tensor_54 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_52, %aten_sigmoid_default_50), kwargs = {})
    %aten_convolution_default_53 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_54, %p_model_22_cv2_1_2_weight, %p_model_22_cv2_1_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_convolution_default_54 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_43, %p_model_22_cv3_1_0_conv_weight, %p_model_22_cv3_1_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_51 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_54,), kwargs = {})
    %aten_mul_tensor_55 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_54, %aten_sigmoid_default_51), kwargs = {})
    %aten_convolution_default_55 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_55, %p_model_22_cv3_1_1_conv_weight, %p_model_22_cv3_1_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_52 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_55,), kwargs = {})
    %aten_mul_tensor_56 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_55, %aten_sigmoid_default_52), kwargs = {})
    %aten_convolution_default_56 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_56, %p_model_22_cv3_1_2_weight, %p_model_22_cv3_1_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_cat_default_14 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_convolution_default_53, %aten_convolution_default_56], 1), kwargs = {})
    %aten_convolution_default_57 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_48, %p_model_22_cv2_2_0_conv_weight, %p_model_22_cv2_2_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_53 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_57,), kwargs = {})
    %aten_mul_tensor_57 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_57, %aten_sigmoid_default_53), kwargs = {})
    %aten_convolution_default_58 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_57, %p_model_22_cv2_2_1_conv_weight, %p_model_22_cv2_2_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_54 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_58,), kwargs = {})
    %aten_mul_tensor_58 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_58, %aten_sigmoid_default_54), kwargs = {})
    %aten_convolution_default_59 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_58, %p_model_22_cv2_2_2_weight, %p_model_22_cv2_2_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_convolution_default_60 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_48, %p_model_22_cv3_2_0_conv_weight, %p_model_22_cv3_2_0_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_55 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_60,), kwargs = {})
    %aten_mul_tensor_59 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_60, %aten_sigmoid_default_55), kwargs = {})
    %aten_convolution_default_61 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_59, %p_model_22_cv3_2_1_conv_weight, %p_model_22_cv3_2_1_conv_bias, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_sigmoid_default_56 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%aten_convolution_default_61,), kwargs = {})
    %aten_mul_tensor_60 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_convolution_default_61, %aten_sigmoid_default_56), kwargs = {})
    %aten_convolution_default_62 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten_mul_tensor_60, %p_model_22_cv3_2_2_weight, %p_model_22_cv3_2_2_bias, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_cat_default_15 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_convolution_default_59, %aten_convolution_default_62], 1), kwargs = {})
    %aten_view_copy_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.view_copy.default](args = (%aten_cat_default_13, [1, 144, -1]), kwargs = {})
    %aten_view_copy_default_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.view_copy.default](args = (%aten_cat_default_14, [1, 144, -1]), kwargs = {})
    %aten_view_copy_default_2 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.view_copy.default](args = (%aten_cat_default_15, [1, 144, -1]), kwargs = {})
    %aten_cat_default_16 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_view_copy_default, %aten_view_copy_default_1, %aten_view_copy_default_2], 2), kwargs = {})
    %aten_split_with_sizes_copy_default_16 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_cat_default_16, [64, 80], 1), kwargs = {})
    %getitem_19 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_16, 0), kwargs = {})
    %getitem_20 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_16, 1), kwargs = {})
    %aten_view_copy_default_3 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.view_copy.default](args = (%getitem_19, [1, 4, 16, 8400]), kwargs = {})
    %aten_permute_copy_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.permute_copy.default](args = (%aten_view_copy_default_3, [0, 2, 1, 3]), kwargs = {})
    %aten__softmax_default : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten._softmax.default](args = (%aten_permute_copy_default, 1, False), kwargs = {})
    %aten_convolution_default_63 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.convolution.default](args = (%aten__softmax_default, %p_model_22_dfl_conv_weight, None, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
    %aten_view_copy_default_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.view_copy.default](args = (%aten_convolution_default_63, [1, 4, 8400]), kwargs = {})
    %aten_split_with_sizes_copy_default_17 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.split_with_sizes_copy.default](args = (%aten_view_copy_default_4, [2, 2], 1), kwargs = {})
    %getitem_21 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_17, 0), kwargs = {})
    %getitem_22 : [num_users=1] = call_function[target=operator.getitem](args = (%aten_split_with_sizes_copy_default_17, 1), kwargs = {})
    %aten_sub_tensor : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.sub.Tensor](args = (%c__prop_tensor_constant2, %getitem_21), kwargs = {})
    %aten_add_tensor_10 : [num_users=2] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%c__prop_tensor_constant2, %getitem_22), kwargs = {})
    %aten_add_tensor_11 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.add.Tensor](args = (%aten_sub_tensor, %aten_add_tensor_10), kwargs = {})
    %dim_order_ops__to_dim_order_copy_default_4 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.dim_order_ops._to_dim_order_copy.default](args = (%_lifted_tensor_constant8,), kwargs = {dtype: torch.float32, dim_order: []})
    %aten_div_tensor : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.div.Tensor](args = (%aten_add_tensor_11, %dim_order_ops__to_dim_order_copy_default_4), kwargs = {})
    %aten_sub_tensor_1 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sub.Tensor](args = (%aten_add_tensor_10, %aten_sub_tensor), kwargs = {})
    %aten_cat_default_17 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_div_tensor, %aten_sub_tensor_1], 1), kwargs = {})
    %aten_mul_tensor_61 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.mul.Tensor](args = (%aten_cat_default_17, %c_model_22_strides), kwargs = {})
    %aten_sigmoid_default_57 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.sigmoid.default](args = (%getitem_20,), kwargs = {})
    %aten_cat_default_18 : [num_users=1] = call_function[target=executorch.exir.dialects.edge._ops.aten.cat.default](args = ([%aten_mul_tensor_61, %aten_sigmoid_default_57], 1), kwargs = {})
    return (aten_cat_default_18,)
ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_weight'), target='model.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_0_conv_bias'), target='model.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_weight'), target='model.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_1_conv_bias'), target='model.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_weight'), target='model.2.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv1_conv_bias'), target='model.2.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_weight'), target='model.2.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_cv2_conv_bias'), target='model.2.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_weight'), target='model.2.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv1_conv_bias'), target='model.2.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_weight'), target='model.2.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_2_m_0_cv2_conv_bias'), target='model.2.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_weight'), target='model.3.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_3_conv_bias'), target='model.3.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_weight'), target='model.4.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv1_conv_bias'), target='model.4.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_weight'), target='model.4.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_cv2_conv_bias'), target='model.4.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_weight'), target='model.4.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv1_conv_bias'), target='model.4.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_weight'), target='model.4.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_0_cv2_conv_bias'), target='model.4.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_weight'), target='model.4.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv1_conv_bias'), target='model.4.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_weight'), target='model.4.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_4_m_1_cv2_conv_bias'), target='model.4.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_weight'), target='model.5.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_5_conv_bias'), target='model.5.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_weight'), target='model.6.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv1_conv_bias'), target='model.6.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_weight'), target='model.6.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_cv2_conv_bias'), target='model.6.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_weight'), target='model.6.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv1_conv_bias'), target='model.6.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_weight'), target='model.6.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_0_cv2_conv_bias'), target='model.6.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_weight'), target='model.6.m.1.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv1_conv_bias'), target='model.6.m.1.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_weight'), target='model.6.m.1.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_6_m_1_cv2_conv_bias'), target='model.6.m.1.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_weight'), target='model.7.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_7_conv_bias'), target='model.7.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_weight'), target='model.8.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv1_conv_bias'), target='model.8.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_weight'), target='model.8.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_cv2_conv_bias'), target='model.8.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_weight'), target='model.8.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv1_conv_bias'), target='model.8.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_weight'), target='model.8.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_8_m_0_cv2_conv_bias'), target='model.8.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_weight'), target='model.9.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv1_conv_bias'), target='model.9.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_weight'), target='model.9.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_9_cv2_conv_bias'), target='model.9.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_weight'), target='model.12.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv1_conv_bias'), target='model.12.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_weight'), target='model.12.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_cv2_conv_bias'), target='model.12.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_weight'), target='model.12.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv1_conv_bias'), target='model.12.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_weight'), target='model.12.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_12_m_0_cv2_conv_bias'), target='model.12.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_weight'), target='model.15.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv1_conv_bias'), target='model.15.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_weight'), target='model.15.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_cv2_conv_bias'), target='model.15.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_weight'), target='model.15.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv1_conv_bias'), target='model.15.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_weight'), target='model.15.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_15_m_0_cv2_conv_bias'), target='model.15.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_weight'), target='model.16.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_16_conv_bias'), target='model.16.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_weight'), target='model.18.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv1_conv_bias'), target='model.18.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_weight'), target='model.18.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_cv2_conv_bias'), target='model.18.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_weight'), target='model.18.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv1_conv_bias'), target='model.18.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_weight'), target='model.18.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_18_m_0_cv2_conv_bias'), target='model.18.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_weight'), target='model.19.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_19_conv_bias'), target='model.19.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_weight'), target='model.21.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv1_conv_bias'), target='model.21.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_weight'), target='model.21.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_cv2_conv_bias'), target='model.21.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_weight'), target='model.21.m.0.cv1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv1_conv_bias'), target='model.21.m.0.cv1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_weight'), target='model.21.m.0.cv2.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_21_m_0_cv2_conv_bias'), target='model.21.m.0.cv2.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_weight'), target='model.22.cv2.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_0_conv_bias'), target='model.22.cv2.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_weight'), target='model.22.cv2.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_1_conv_bias'), target='model.22.cv2.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_weight'), target='model.22.cv2.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_0_2_bias'), target='model.22.cv2.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_weight'), target='model.22.cv2.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_0_conv_bias'), target='model.22.cv2.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_weight'), target='model.22.cv2.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_1_conv_bias'), target='model.22.cv2.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_weight'), target='model.22.cv2.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_1_2_bias'), target='model.22.cv2.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_weight'), target='model.22.cv2.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_0_conv_bias'), target='model.22.cv2.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_weight'), target='model.22.cv2.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_1_conv_bias'), target='model.22.cv2.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_weight'), target='model.22.cv2.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv2_2_2_bias'), target='model.22.cv2.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_weight'), target='model.22.cv3.0.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_0_conv_bias'), target='model.22.cv3.0.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_weight'), target='model.22.cv3.0.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_1_conv_bias'), target='model.22.cv3.0.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_weight'), target='model.22.cv3.0.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_0_2_bias'), target='model.22.cv3.0.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_weight'), target='model.22.cv3.1.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_0_conv_bias'), target='model.22.cv3.1.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_weight'), target='model.22.cv3.1.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_1_conv_bias'), target='model.22.cv3.1.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_weight'), target='model.22.cv3.1.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_1_2_bias'), target='model.22.cv3.1.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_weight'), target='model.22.cv3.2.0.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_0_conv_bias'), target='model.22.cv3.2.0.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_weight'), target='model.22.cv3.2.1.conv.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_1_conv_bias'), target='model.22.cv3.2.1.conv.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_weight'), target='model.22.cv3.2.2.weight', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_cv3_2_2_bias'), target='model.22.cv3.2.2.bias', persistent=None), InputSpec(kind=<InputKind.PARAMETER: 2>, arg=TensorArgument(name='p_model_22_dfl_conv_weight'), target='model.22.dfl.conv.weight', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c__prop_tensor_constant2'), target='_prop_tensor_constant2', persistent=None), InputSpec(kind=<InputKind.CONSTANT_TENSOR: 4>, arg=TensorArgument(name='c_model_22_strides'), target='model.22.strides', persistent=None), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant0'), target='_lifted_tensor_constant0', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant1'), target='_lifted_tensor_constant1', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant2'), target='_lifted_tensor_constant2', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant3'), target='_lifted_tensor_constant3', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant4'), target='_lifted_tensor_constant4', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant5'), target='_lifted_tensor_constant5', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant6'), target='_lifted_tensor_constant6', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant7'), target='_lifted_tensor_constant7', persistent=True), InputSpec(kind=<InputKind.BUFFER: 3>, arg=TensorArgument(name='_lifted_tensor_constant8'), target='_lifted_tensor_constant8', persistent=True), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='x'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='aten_cat_default_18'), target=None)])

@github-actions github-actions bot mentioned this issue Mar 10, 2025
Open
@JacobSzwejbka
Copy link
Contributor

Oh wait the error happens after execute? I think that actually means its likely its failing inside an op somewhere. Sorry I missed that. cc @Gasoonjia do you know an easy way for someone in OSS to debug what operator is failing in mid execution? I wouldve thought some logging should appear.

@Gasoonjia
Copy link
Contributor

I don't think we have any specific tool to show operator failing; if ET_LOG didn't work as expected, perhaps sanity checks of some operators don't cover enough?

This was referenced Mar 17, 2025
Open
Open
@corehalt
Copy link
Author

@JacobSzwejbka @iseeyuan any updates on this issue?

@github-actions github-actions bot mentioned this issue Mar 31, 2025
Open
@github-actions github-actions bot mentioned this issue Apr 7, 2025
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@JacobSzwejbka JacobSzwejbka

Labels
module: runtime Issues related to the core runtime and code under runtime/ triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
ExecuTorch Core
Status: In progress
Milestone
No milestone
Development

No branches or pull requests
4 participants
@iseeyuan @JacobSzwejbka @Gasoonjia @corehalt