🐛 [Bug] RuntimeError: Method 'forward' is not defined.

[zhenglongjiepheonix]

Bug Description

To Reproduce

import torch
from torchvision.models.detection import fasterrcnn_resnet50_fpn
import trtorch

model = fasterrcnn_resnet50_fpn()
scripted = torch.jit.script(model)
trt = trtorch.compile(scripted, {"input_shapes": [(1, 3, 224, 224)]})# error occurs!

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/zhenglongjie/miniconda3/envs/mmlab/lib/python3.7/site-packages/trtorch/_compiler.py", line 81, in compile
    compiled_cpp_mod = trtorch._C.compile_graph(module._c, _parse_compile_spec(compile_spec))
RuntimeError: Method 'forward' is not defined.

Expected behavior

no error should occur

Environment

Build information about the TRTorch compiler can be found by turning on debug messages

• TRTorch Version (e.g. 0.2.0): 0.5.0a
• PyTorch Version (e.g. 1.0):1.9
• CPU Architecture:x86_64
• OS (e.g., Linux):ubuntu 18.04
• How you installed PyTorch (conda, pip, libtorch, source):pip
• Build command you used (if compiling from source):
• Are you using local sources or building from archives:
• Python version:3.7
• CUDA version:11.1
• GPU models and configuration:
• Any other relevant information:

Additional context

[zhenglongjiepheonix]

when I try alexnet it works just fine

[p1x31]

This model doesn't support tracing and has a complex control flow for multiple heads RPNHead, TwoMLPHead, RoIHeads etc
Forward method in every class there

[github-actions]

This issue has not seen activity for 90 days, Remove stale label or comment or this will be closed in 10 days