add return value in deploy.py

mmdeploy/apis/pytorch2onnx.py

@@ -3,6 +3,7 @@ from typing import Any, Optional, Union
 
 import mmcv
 import torch
+import torch.multiprocessing as mp
 
 from mmdeploy.utils import (RewriterContext, patch_model,
                             register_extra_symbolics)
@@ -14,7 +15,10 @@ def torch2onnx(img: Any,
                deploy_cfg: Union[str, mmcv.Config],
                model_cfg: Union[str, mmcv.Config],
                model_checkpoint: Optional[str] = None,
-               device: str = 'cuda:0'):
+               device: str = 'cuda:0',
+               ret_value: Optional[mp.Value] = None):
+    ret_value.value = -1
+
     # load deploy_cfg if needed
     if isinstance(deploy_cfg, str):
         deploy_cfg = mmcv.Config.fromfile(deploy_cfg)
@@ -40,8 +44,7 @@ def torch2onnx(img: Any,
 
     torch_model = init_model(codebase, model_cfg, model_checkpoint, device)
     data, model_inputs = create_input(codebase, model_cfg, img, device)
-    patched_model = patch_model(
+    patched_model = patch_model(torch_model, cfg=deploy_cfg, backend=backend)
-        torch_model, cfg=deploy_cfg, backend=backend, data=data)
 
     if not isinstance(model_inputs, torch.Tensor):
         model_inputs = model_inputs[0]
@@ -57,3 +60,5 @@ def torch2onnx(img: Any,
             dynamic_axes=pytorch2onnx_cfg.get('dynamic_axes', None),
             keep_initializers_as_inputs=pytorch2onnx_cfg[
                 'keep_initializers_as_inputs'])
+
+    ret_value.value = 0

mmdeploy/utils/module_rewriter.py

@@ -70,10 +70,11 @@ MODULE_REWRITERS = RewriteModuleRegistry(
 def patch_model(model: nn.Module,
                 cfg: Dict,
                 backend: str = 'default',
+                recursive: bool = True,
                 **kwargs) -> nn.Module:
 
     def _patch_impl(model, cfg, **kwargs):
-        if hasattr(model, 'named_children'):
+        if recursive and hasattr(model, 'named_children'):
             for name, module in model.named_children():
                 model._modules[name] = _patch_impl(module, cfg, **kwargs)
         return MODULE_REWRITERS.build(

tools/deploy.py

@@ -3,6 +3,7 @@ import logging
 import os.path as osp
 
 import mmcv
+import torch.multiprocessing as mp
 from torch.multiprocessing import Process, set_start_method
 
 from mmdeploy.apis import torch2onnx
@@ -34,15 +35,23 @@ def main():
     # create work_dir if not
     mmcv.mkdir_or_exist(osp.abspath(args.work_dir))
 
+    ret_value = mp.Value('d', 0, lock=False)
+
     # convert model
     logging.info('start torch2onnx conversion.')
     process = Process(
         target=torch2onnx,
         args=(args.img, args.work_dir, deploy_cfg, model_cfg, checkpoint),
-        kwargs=dict(device=args.device))
+        kwargs=dict(device=args.device, ret_value=ret_value))
     process.start()
     process.join()
 
+    if ret_value.value != 0:
+        logging.error('torch2onnx failed.')
+        exit()
+    else:
+        logging.info('torch2onnx success.')
+
 
 if __name__ == '__main__':
     main()