Co-DETR
Co-DETR copied to clipboard
Sense-X
pytorch to onnx
/home/aigroup/chenzx/ws_internImage/bin/python3.8 /home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/init.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
warnings.warn(
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py:284: UserWarning: Arguments like --mean, --std, --dataset would be parsed directly from config file and are deprecated and will be removed in future releases.
warnings.warn('Arguments like --mean, --std, --dataset would be
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/onnx/symbolic.py:481: UserWarning: DeprecationWarning: This function will be deprecated in future. Welcome to use the unified model deployment toolbox MMDeploy: https://github.com/open-mmlab/mmdeploy
warnings.warn(msg)
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
2023-08-08 16:30:39,106 - mmcv - INFO - initialize RPNHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01}
2023-08-08 16:30:39,108 - mmcv - INFO -
rpn_conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,108 - mmcv - INFO - rpn_conv.bias - torch.Size([256]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,108 - mmcv - INFO - rpn_cls.weight - torch.Size([9, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,108 - mmcv - INFO - rpn_cls.bias - torch.Size([9]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,108 - mmcv - INFO - rpn_reg.weight - torch.Size([36, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,108 - mmcv - INFO - rpn_reg.bias - torch.Size([36]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,185 - mmcv - INFO - initialize Shared2FCBBoxHead with init_cfg [{'type': 'Normal', 'std': 0.01, 'override': {'name': 'fc_cls'}}, {'type': 'Normal', 'std': 0.001, 'override': {'name': 'fc_reg'}}, {'type': 'Xavier', 'distribution': 'uniform', 'override': [{'name': 'shared_fcs'}, {'name': 'cls_fcs'}, {'name': 'reg_fcs'}]}] 2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.fc_cls.weight - torch.Size([81, 1024]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.fc_cls.bias - torch.Size([81]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.fc_reg.weight - torch.Size([320, 1024]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.fc_reg.bias - torch.Size([320]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.shared_fcs.0.weight - torch.Size([1024, 12544]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.shared_fcs.0.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.shared_fcs.1.weight - torch.Size([1024, 1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 16:30:39,236 - mmcv - INFO - bbox_head.shared_fcs.1.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/anchor_head.py:116: UserWarning: DeprecationWarning: num_anchors is deprecated, for consistency or also use num_base_priors instead
warnings.warn('DeprecationWarning: num_anchors is deprecated, '
2023-08-08 16:30:39,248 - mmcv - INFO - initialize CoATSSHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01, 'override': {'type': 'Normal', 'name': 'atss_cls', 'std': 0.01, 'bias_prob': 0.01}}
2023-08-08 16:30:39,255 - mmcv - INFO -
cls_convs.0.conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO -
cls_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
cls_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO - reg_convs.0.conv.weight - torch.Size([256, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO -
reg_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
reg_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO - atss_cls.weight - torch.Size([80, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 16:30:39,255 - mmcv - INFO - atss_cls.bias - torch.Size([80]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 16:30:39,255 - mmcv - INFO - atss_reg.weight - torch.Size([4, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO - atss_reg.bias - torch.Size([4]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO - atss_centerness.weight - torch.Size([1, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO - atss_centerness.bias - torch.Size([1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.0.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.1.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.2.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.3.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.4.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 16:30:39,255 - mmcv - INFO -
scales.5.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
load checkpoint from local path: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/model/co_dino_5scale_swin_large_3x_coco.pth /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:423: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if W % self.patch_size[1] != 0: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:425: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if H % self.patch_size[0] != 0: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:362: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! Hp = int(np.ceil(H / self.window_size)) * self.window_size /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:363: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! Wp = int(np.ceil(W / self.window_size)) * self.window_size /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:203: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! assert L == H * W, "input feature has wrong size" /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:66: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! B = int(windows.shape[0] / (H * W / window_size / window_size)) /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:241: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if pad_r > 0 or pad_b > 0: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:272: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! assert L == H * W, "input feature has wrong size" /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:277: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! pad_input = (H % 2 == 1) or (W % 2 == 1) /home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:278: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if pad_input: ============= Diagnostic Run torch.onnx.export version 2.0.0+cu117 ============= verbose: False, log level: Level.ERROR ======================= 0 NONE 0 NOTE 0 WARNING 0 ERROR ========================
Traceback (most recent call last):
File "/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py", line 320, in
I have modified the
onnx_exportfunction, you can try it again
the problem is solved，there is a bug
/home/aigroup/chenzx/ws_internImage/bin/python3.8 /home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/init.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
warnings.warn(
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py:284: UserWarning: Arguments like --mean, --std, --dataset would be parsed directly from config file and are deprecated and will be removed in future releases.
warnings.warn('Arguments like --mean, --std, --dataset would be
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/onnx/symbolic.py:481: UserWarning: DeprecationWarning: This function will be deprecated in future. Welcome to use the unified model deployment toolbox MMDeploy: https://github.com/open-mmlab/mmdeploy
warnings.warn(msg)
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
2023-08-08 17:01:24,154 - mmcv - INFO - initialize RPNHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01}
2023-08-08 17:01:24,156 - mmcv - INFO -
rpn_conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,156 - mmcv - INFO - rpn_conv.bias - torch.Size([256]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,156 - mmcv - INFO - rpn_cls.weight - torch.Size([9, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,156 - mmcv - INFO - rpn_cls.bias - torch.Size([9]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,156 - mmcv - INFO - rpn_reg.weight - torch.Size([36, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,156 - mmcv - INFO - rpn_reg.bias - torch.Size([36]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,232 - mmcv - INFO - initialize Shared2FCBBoxHead with init_cfg [{'type': 'Normal', 'std': 0.01, 'override': {'name': 'fc_cls'}}, {'type': 'Normal', 'std': 0.001, 'override': {'name': 'fc_reg'}}, {'type': 'Xavier', 'distribution': 'uniform', 'override': [{'name': 'shared_fcs'}, {'name': 'cls_fcs'}, {'name': 'reg_fcs'}]}] 2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.fc_cls.weight - torch.Size([81, 1024]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.fc_cls.bias - torch.Size([81]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.fc_reg.weight - torch.Size([320, 1024]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.fc_reg.bias - torch.Size([320]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.shared_fcs.0.weight - torch.Size([1024, 12544]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.shared_fcs.0.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.shared_fcs.1.weight - torch.Size([1024, 1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:01:24,283 - mmcv - INFO - bbox_head.shared_fcs.1.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/anchor_head.py:116: UserWarning: DeprecationWarning: num_anchors is deprecated, for consistency or also use num_base_priors instead
warnings.warn('DeprecationWarning: num_anchors is deprecated, '
2023-08-08 17:01:24,294 - mmcv - INFO - initialize CoATSSHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01, 'override': {'type': 'Normal', 'name': 'atss_cls', 'std': 0.01, 'bias_prob': 0.01}}
2023-08-08 17:01:24,299 - mmcv - INFO -
cls_convs.0.conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO -
cls_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
cls_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO - reg_convs.0.conv.weight - torch.Size([256, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO -
reg_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
reg_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO - atss_cls.weight - torch.Size([80, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 17:01:24,299 - mmcv - INFO - atss_cls.bias - torch.Size([80]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 17:01:24,299 - mmcv - INFO - atss_reg.weight - torch.Size([4, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO - atss_reg.bias - torch.Size([4]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO - atss_centerness.weight - torch.Size([1, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO - atss_centerness.bias - torch.Size([1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.0.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.1.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.2.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.3.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.4.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:01:24,299 - mmcv - INFO -
scales.5.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
load checkpoint from local path: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/model/co_dino_5scale_swin_large_3x_coco.pth
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:423: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if W % self.patch_size[1] != 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:425: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if H % self.patch_size[0] != 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:362: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
Hp = int(np.ceil(H / self.window_size)) * self.window_size
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:363: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
Wp = int(np.ceil(W / self.window_size)) * self.window_size
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:203: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert L == H * W, "input feature has wrong size"
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:66: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
B = int(windows.shape[0] / (H * W / window_size / window_size))
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:241: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if pad_r > 0 or pad_b > 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:272: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert L == H * W, "input feature has wrong size"
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:277: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
pad_input = (H % 2 == 1) or (W % 2 == 1)
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:278: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if pad_input:
Traceback (most recent call last):
File "/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py", line 320, in
I fixed it like this
here is my new problem
/home/aigroup/chenzx/ws_internImage/bin/python3.8 /home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/init.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
warnings.warn(
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py:284: UserWarning: Arguments like --mean, --std, --dataset would be parsed directly from config file and are deprecated and will be removed in future releases.
warnings.warn('Arguments like --mean, --std, --dataset would be
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/onnx/symbolic.py:481: UserWarning: DeprecationWarning: This function will be deprecated in future. Welcome to use the unified model deployment toolbox MMDeploy: https://github.com/open-mmlab/mmdeploy
warnings.warn(msg)
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
2023-08-08 17:02:19,967 - mmcv - INFO - initialize RPNHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01}
2023-08-08 17:02:19,969 - mmcv - INFO -
rpn_conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:19,969 - mmcv - INFO - rpn_conv.bias - torch.Size([256]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:19,969 - mmcv - INFO - rpn_cls.weight - torch.Size([9, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:19,969 - mmcv - INFO - rpn_cls.bias - torch.Size([9]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:19,969 - mmcv - INFO - rpn_reg.weight - torch.Size([36, 256, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:19,969 - mmcv - INFO - rpn_reg.bias - torch.Size([36]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,044 - mmcv - INFO - initialize Shared2FCBBoxHead with init_cfg [{'type': 'Normal', 'std': 0.01, 'override': {'name': 'fc_cls'}}, {'type': 'Normal', 'std': 0.001, 'override': {'name': 'fc_reg'}}, {'type': 'Xavier', 'distribution': 'uniform', 'override': [{'name': 'shared_fcs'}, {'name': 'cls_fcs'}, {'name': 'reg_fcs'}]}] 2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.fc_cls.weight - torch.Size([81, 1024]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.fc_cls.bias - torch.Size([81]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.fc_reg.weight - torch.Size([320, 1024]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.fc_reg.bias - torch.Size([320]): NormalInit: mean=0, std=0.001, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.shared_fcs.0.weight - torch.Size([1024, 12544]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.shared_fcs.0.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.shared_fcs.1.weight - torch.Size([1024, 1024]): XavierInit: gain=1, distribution=uniform, bias=0
2023-08-08 17:02:20,095 - mmcv - INFO - bbox_head.shared_fcs.1.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/anchor_head.py:116: UserWarning: DeprecationWarning: num_anchors is deprecated, for consistency or also use num_base_priors instead
warnings.warn('DeprecationWarning: num_anchors is deprecated, '
2023-08-08 17:02:20,106 - mmcv - INFO - initialize CoATSSHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01, 'override': {'type': 'Normal', 'name': 'atss_cls', 'std': 0.01, 'bias_prob': 0.01}}
2023-08-08 17:02:20,111 - mmcv - INFO -
cls_convs.0.conv.weight - torch.Size([256, 256, 3, 3]):
NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO -
cls_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,111 - mmcv - INFO -
cls_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,111 - mmcv - INFO - reg_convs.0.conv.weight - torch.Size([256, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO -
reg_convs.0.gn.weight - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,111 - mmcv - INFO -
reg_convs.0.gn.bias - torch.Size([256]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,111 - mmcv - INFO - atss_cls.weight - torch.Size([80, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 17:02:20,111 - mmcv - INFO - atss_cls.bias - torch.Size([80]): NormalInit: mean=0, std=0.01, bias=-4.59511985013459
2023-08-08 17:02:20,111 - mmcv - INFO - atss_reg.weight - torch.Size([4, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO - atss_reg.bias - torch.Size([4]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO - atss_centerness.weight - torch.Size([1, 256, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO - atss_centerness.bias - torch.Size([1]): NormalInit: mean=0, std=0.01, bias=0
2023-08-08 17:02:20,111 - mmcv - INFO -
scales.0.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,111 - mmcv - INFO -
scales.1.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,112 - mmcv - INFO -
scales.2.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,112 - mmcv - INFO -
scales.3.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,112 - mmcv - INFO -
scales.4.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
2023-08-08 17:02:20,112 - mmcv - INFO -
scales.5.scale - torch.Size([]):
The value is the same before and after calling init_weights of CoATSSHead
load checkpoint from local path: /home/aigroup/chenzx/ws_internImage/code/Co-DETR/model/co_dino_5scale_swin_large_3x_coco.pth The model and loaded state dict do not match exactly
missing keys in source state_dict: query_head.input_proj.weight, query_head.input_proj.bias, query_head.fc_cls.weight, query_head.fc_cls.bias, query_head.reg_ffn.layers.0.0.weight, query_head.reg_ffn.layers.0.0.bias, query_head.reg_ffn.layers.1.weight, query_head.reg_ffn.layers.1.bias, query_head.fc_reg.weight, query_head.fc_reg.bias
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:423: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if W % self.patch_size[1] != 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:425: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if H % self.patch_size[0] != 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:362: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
Hp = int(np.ceil(H / self.window_size)) * self.window_size
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:363: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
Wp = int(np.ceil(W / self.window_size)) * self.window_size
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:203: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert L == H * W, "input feature has wrong size"
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:66: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
B = int(windows.shape[0] / (H * W / window_size / window_size))
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:241: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if pad_r > 0 or pad_b > 0:
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:272: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert L == H * W, "input feature has wrong size"
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:277: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
pad_input = (H % 2 == 1) or (W % 2 == 1)
/home/aigroup/chenzx/ws_internImage/code/Co-DETR/mmdet/models/dense_heads/swin_transformer.py:278: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if pad_input:
/home/aigroup/chenzx/ws_internImage/lib/python3.8/site-packages/mmcv/cnn/bricks/wrappers.py:45: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if x.numel() == 0 and obsolete_torch_version(TORCH_VERSION, (1, 4)):
Traceback (most recent call last):
File "/home/aigroup/chenzx/ws_internImage/code/Co-DETR/tools/deployment/pytorch2onnx.py", line 320, in
Is this problem solved?
没有解决，现在看到工程项目又更新了，我找个时间更新项目再试试
@TempleX98 更新了以后还是有这个问题 TAT
return F.conv2d(input, weight, bias, self.stride,
RuntimeError: Given groups=1, weight of size [256, 2048, 1, 1], expected input[1, 256, 200, 304] to have 2048 channels, but got 256 channels instead
@xueyingliu, @chenzx2, @jielanZhang, Hi, I am sorry that there are some unsolved torch2onnx issues. Our repo is implemented using an older version of mmdet v2.25, which no longer maintains the feature of model export. Co-DETR has been recently incorporated into the official mmdet v3.x repo, and you can use this official implementation as well as MMDeploy for model export.
@TempleX98 I have tried your suggestion but I encounter some problems but when verify the documentation in mmdeploy, I found that CO-DETR is not supported, here is the list of the supported models https://github.com/open-mmlab/mmdeploy/blob/main/docs/en/03-benchmark/supported_models.md
@MarouaneMja Hi, I find this PR (https://github.com/open-mmlab/mmdetection/pull/10910) of mmdet v3.x supports onnx export and hope this can help you.
@Mayyyybe The inference architecture of Co-DINO is the same as DINO. And MMdeploy supports the model export of DINO method.
Hi @TempleX98 , I managed to export the CO-DETR to onnx using mmdeploy as you suggested, however the SoftNonMaxSuppression is not supported by onnxruntime failed:Fatal error: SoftNonMaxSuppression is not a registered function/op, do you have any suggestion how to bypass this problem?
Hi @TempleX98 , I managed to export the CO-DETR to onnx using mmdeploy as you suggested, however the SoftNonMaxSuppression is not supported by onnxruntime
failed:Fatal error: SoftNonMaxSuppression is not a registered function/op, do you have any suggestion how to bypass this problem?
Just remove the NMS operation in the config.
Thank you for your help @TempleX98 , it worked. However when I launch infernce with triton inference server , the onnx model takes more space in gpu memory than using a simple Python backend which very stange lolll
嗨，我设法按照您的建议使用 mmdeploy 将 CO-DETR 导出到 onnx，但是 onnxruntime 不支持 SoftNonMaxSuppression ，您有什么建议如何绕过这个问题吗？
failed:Fatal error: SoftNonMaxSuppression is not a registered function/opHello, I am also trying to export CO-DETR to onnx using mmdeploy. I would like to ask which model_cfg you are using is“https://github.com/RunningLeon/mmdetection/blob/support_dino_onnx/projects/CO-DETR/configs/codino/co_dino_5scale_swin_l_16xb1_16e_o365tococo.py “ or ”https://github.com/Sense-X/Co-DETR/blob/main/projects/configs/co_dino/co_dino_5scale_swin_large_16e_o365tococo.py“
And could you please share the deploy_cfg for me to learn it.I would be very grateful if you could help me!
@xinlin-xiao I am using the first config https://github.com/RunningLeon/mmdetection/blob/support_dino_onnx/projects/CO-DETR/configs/codino/co_dino_5scale_swin_l_16xb1_16e_o365tococo.py “, for the deploy_config I am using "detection_onnxruntime_dynamic.py" in mmdeploy
@xinlin-xiao I am using the first config https://github.com/RunningLeon/mmdetection/blob/support_dino_onnx/projects/CO-DETR/configs/codino/co_dino_5scale_swin_l_16xb1_16e_o365tococo.py “, for the deploy_config I am using "detection_onnxruntime_dynamic.py" in mmdeploy
Thank you for your help!But,I trying to export CO-DETR to onnx using mmdeploy in this:
python3 /root/workspace/mmdeploy/tools/deploy.py \ /root/workspace/mmdeploy/configs/mmdet/detection/detection_onnxruntime_dynamic.py\ /mnt/data/train-yolov7-images/pytorch2onn-mmdeploy/co_dino-new/co_dino_5scale_swin_l_16xb1_16e_o365tococo.py \ /mnt/data/train-yolov7-images/pytorch2onn-mmdeploy/co_dino_5scale_swin_large_16e_o365tococo.pth\ /root/workspace/mmdetection-support_dino_onnx/demo/demo.jpg \ --work-dir /mnt/data/train-yolov7-images/pytorch2onn-mmdeploy \ --device cuda \ --dump-info
It reported an error：
`============= Diagnostic Run torch.onnx.export version 2.0.0+cu118 =============
verbose: False, log level: Level.ERROR
======================= 0 NONE 0 NOTE 0 WARNING 0 ERROR ========================
Process Process-2:
Traceback (most recent call last):
File "/usr/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
self.run()
File "/usr/lib/python3.8/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 107, in call
ret = func(*args, **kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/pytorch2onnx.py", line 98, in torch2onnx
export(
File "/root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 356, in wrap
return self.call_function(func_name, *args, **kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 326, in call_function
return self.call_function_local(func_name, *args, **kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 275, in call_function_local
return pipe_caller(*args, **kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 107, in call
ret = func(*args, **kwargs)
File "/root/workspace/mmdeploy/mmdeploy/apis/onnx/export.py", line 138, in export
torch.onnx.export(
File "/usr/local/lib/python3.8/dist-packages/torch/onnx/utils.py", line 506, in export
_export(
File "/usr/local/lib/python3.8/dist-packages/torch/onnx/utils.py", line 1548, in _export
graph, params_dict, torch_out = _model_to_graph(
File "/root/workspace/mmdeploy/mmdeploy/apis/onnx/optimizer.py", line 27, in model_to_graph__custom_optimizer
graph, params_dict, torch_out = ctx.origin_func(*args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/torch/onnx/utils.py", line 1113, in _model_to_graph
graph, params, torch_out, module = _create_jit_graph(model, args)
File "/usr/local/lib/python3.8/dist-packages/torch/onnx/utils.py", line 989, in _create_jit_graph
graph, torch_out = _trace_and_get_graph_from_model(model, args)
File "/usr/local/lib/python3.8/dist-packages/torch/onnx/utils.py", line 893, in _trace_and_get_graph_from_model
trace_graph, torch_out, inputs_states = torch.jit._get_trace_graph(
File "/usr/local/lib/python3.8/dist-packages/torch/jit/_trace.py", line 1268, in _get_trace_graph
outs = ONNXTracedModule(f, strict, _force_outplace, return_inputs, _return_inputs_states)(*args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/torch/jit/_trace.py", line 127, in forward
graph, out = torch._C._create_graph_by_tracing(
File "/usr/local/lib/python3.8/dist-packages/torch/jit/_trace.py", line 121, in wrapper
out_vars, _ = _flatten(outs)
RuntimeError: Only tuples, lists and Variables are supported as JIT inputs/outputs. Dictionaries and strings are also accepted, but their usage is not recommended. Here, received an input of unsupported type: DetDataSample
11/16 16:55:21 - mmengine - ERROR - /root/workspace/mmdeploy/mmdeploy/apis/core/pipeline_manager.py - pop_mp_output - 80 - mmdeploy.apis.pytorch2onnx.torch2onnx with Call id: 0 failed. exit.
`
The ”co_dino_5scale_swin_large_16e_o365tococo.pth“is in https://github.com/Sense-X/Co-DETR/tree/main ：
Could you please tell me which checkpoint is you use？And I find someone use ”model_checkpoint = 'checkpoints/co_dino_5scale_r50_lsj_8xb2_1x_coco-69a72d67.pth'“ in https://github.com/open-mmlab/mmdetection/issues/11011 It is also return the same error！ So how did you succeed export CO-DETR to onnx using mmdeploy without SoftNonMaxSuppression.I would be very grateful if you could help me!
Just remove the soft_nms from the config, and you can add it later
Do you have any suggestions on how to fix this mistake?
RuntimeError:Only tuples, lists and Variables are supported as JIT inputs/outputs. Dictionaries and strings are also accepted, but their usage is not recommended. Here, received an input of unsupported type: DetDataSample
Yes, DetDataSample are not supported by JIT , you have to convert the final output to a tuple format. Try, to run a simple Detr to get an idea what the output looks like , then do the same
@MarouaneMja
Yes, DetDataSample are not supported by JIT , you have to convert the final output to a tuple format. Try, to run a simple Detr to get an idea what the output looks like , then do the same
I am trying find the model output is : `out:[<DetDataSample(
META INFORMATION
img_path: '/root/workspace/mmdetection-support_dino_onnx/demo/demo.jpg'
scale_factor: (2.9984375, 2.997658079625293)
img_shape: (1280, 1919)
batch_input_shape: (1280, 1919)
ori_shape: (427, 640)
img_id: 0
pad_shape: (1280, 1919)
DATA FIELDS
pred_instances: <InstanceData(
META INFORMATION
DATA FIELDS
scores: tensor([0.1154, 0.1098, 0.1047, 0.0994, 0.0982, 0.0867, 0.0844, 0.0714, 0.0704,
0.0603, 0.0577, 0.0564, 0.0543, 0.0538, 0.0520, 0.0511, 0.0509, 0.0509,
0.0502, 0.0498, 0.0495, 0.0491, 0.0461, 0.0460, 0.0440, 0.0434, 0.0432,
0.0425, 0.0422, 0.0419, 0.0409, 0.0407, 0.0407, 0.0407, 0.0391, 0.0390,
0.0381, 0.0375, 0.0372, 0.0372, 0.0369, 0.0367, 0.0367, 0.0367, 0.0362,
0.0358, 0.0355, 0.0350, 0.0349, 0.0349, 0.0347, 0.0347, 0.0345, 0.0342,
0.0341, 0.0340, 0.0340, 0.0339, 0.0339, 0.0338, 0.0332, 0.0332, 0.0331,
0.0330, 0.0327, 0.0326, 0.0326, 0.0326, 0.0325, 0.0324, 0.0322, 0.0322,
0.0318, 0.0317, 0.0316, 0.0315, 0.0314, 0.0312, 0.0312, 0.0310, 0.0305,
0.0304, 0.0304, 0.0304, 0.0299, 0.0299, 0.0299, 0.0298, 0.0298, 0.0297,
0.0296, 0.0295, 0.0293, 0.0291, 0.0290, 0.0290, 0.0290, 0.0290, 0.0289,
0.0287, 0.0285, 0.0284, 0.0283, 0.0282, 0.0281, 0.0281, 0.0281, 0.0280,
0.0280, 0.0280, 0.0279, 0.0279, 0.0277, 0.0277, 0.0276, 0.0275, 0.0274,
0.0274, 0.0272, 0.0271, 0.0271, 0.0269, 0.0266, 0.0266, 0.0265, 0.0265,
0.0263, 0.0263, 0.0262, 0.0262, 0.0260, 0.0260, 0.0260, 0.0260, 0.0260,
0.0259, 0.0258, 0.0256, 0.0256, 0.0255, 0.0255, 0.0255, 0.0254, 0.0253,
0.0251, 0.0250, 0.0248, 0.0246, 0.0246, 0.0246, 0.0246, 0.0246, 0.0244,
0.0244, 0.0244, 0.0244, 0.0244, 0.0243, 0.0243, 0.0243, 0.0243, 0.0242,
0.0241, 0.0236, 0.0235, 0.0235, 0.0234, 0.0234, 0.0234, 0.0233, 0.0233,
0.0233, 0.0232, 0.0232, 0.0231, 0.0231, 0.0231, 0.0230, 0.0230, 0.0230,
0.0230, 0.0230, 0.0228, 0.0228, 0.0227, 0.0226, 0.0226, 0.0225, 0.0225,
0.0225, 0.0223, 0.0222, 0.0221, 0.0220, 0.0220, 0.0219, 0.0219, 0.0218,
0.0218, 0.0216, 0.0216, 0.0216, 0.0215, 0.0214, 0.0214, 0.0214, 0.0213,
0.0213, 0.0213, 0.0213, 0.0213, 0.0212, 0.0212, 0.0212, 0.0212, 0.0211,
0.0209, 0.0209, 0.0208, 0.0208, 0.0208, 0.0208, 0.0208, 0.0208, 0.0207,
0.0207, 0.0207, 0.0207, 0.0207, 0.0206, 0.0205, 0.0205, 0.0205, 0.0205,
0.0204, 0.0204, 0.0068, 0.0067, 0.0056, 0.0051, 0.0048, 0.0048, 0.0046,
0.0041, 0.0040, 0.0037, 0.0036, 0.0033, 0.0030, 0.0030, 0.0028, 0.0028,
0.0027, 0.0022, 0.0021, 0.0016, 0.0015, 0.0014, 0.0014, 0.0013, 0.0013,
0.0012, 0.0012, 0.0011, 0.0010], device='cuda:0')
bboxes: tensor([[575.6778, 238.4051, 640.0000, 377.5103],
[ 6.4190, 1.7029, 638.2518, 377.3401],
[402.1796, 282.1749, 640.0000, 426.7463],
...,
[ 0.9140, 0.0000, 640.0000, 249.7452],
[ 0.0000, 206.1276, 422.8578, 427.0000],
[ 0.9140, 0.0000, 640.0000, 249.7452]], device='cuda:0')
labels: tensor([39, 0, 0, 34, 0, 0, 69, 60, 41, 72, 79, 39, 71, 34, 39, 60, 60, 39,
39, 34, 0, 0, 0, 56, 39, 41, 43, 71, 34, 44, 39, 34, 69, 0, 0, 39,
41, 17, 39, 76, 0, 0, 34, 39, 79, 79, 0, 72, 12, 39, 39, 34, 39, 79,
44, 39, 39, 44, 0, 39, 41, 41, 34, 44, 45, 0, 78, 39, 79, 60, 34, 0,
41, 32, 39, 69, 39, 71, 0, 79, 0, 39, 0, 44, 55, 0, 56, 39, 34, 78,
71, 4, 44, 38, 34, 72, 43, 60, 72, 0, 12, 34, 39, 39, 45, 0, 0, 62,
61, 0, 0, 61, 34, 34, 69, 39, 72, 79, 0, 34, 0, 78, 67, 0, 72, 34,
41, 39, 0, 70, 0, 71, 40, 0, 41, 44, 60, 0, 79, 0, 39, 0, 38, 34,
0, 71, 79, 39, 0, 0, 0, 29, 45, 0, 53, 34, 12, 0, 60, 67, 79, 34,
0, 0, 78, 39, 34, 45, 43, 40, 44, 6, 62, 79, 0, 78, 0, 0, 67, 71,
68, 0, 59, 32, 41, 0, 0, 39, 44, 12, 56, 44, 39, 41, 39, 39, 12, 56,
34, 69, 71, 40, 15, 38, 79, 60, 44, 72, 79, 44, 56, 44, 0, 40, 12, 72,
39, 60, 39, 0, 79, 0, 0, 44, 79, 72, 38, 79, 40, 41, 34, 13, 43, 79,
41, 79, 34, 60, 39, 0, 0, 0, 39, 44, 0, 34, 39, 39, 0, 39, 0, 0,
39, 34, 69, 0, 34, 41, 60, 0, 79, 44, 71, 60, 17], device='cuda:0')
) at 0x7fb8680766d0>
ignored_instances: <InstanceData(
META INFORMATION
DATA FIELDS
) at 0x7fb82c7c1820>
gt_instances: <InstanceData(
META INFORMATION
DATA FIELDS
) at 0x7fb82c7c1850>
) at 0x7fb82c7c18e0>]
But，I do not know the right output looks like,trying use torch.jit.script to keep only labels, bboxes, tensor, scores in pred_instances is also return errors.
if you succeed export CO-DETR to onnx or TorchScript,could you please share your modified code and the right output to help me fix the errors？
I would be very grateful if you could help me! This problem has been bothering me for a long time.
Just remove the soft_nms from the config, and you can add it later
hello，the config file，I don’t find the soft_nms
Hi! I figured out to convert co-dino to onnx and would like to share with you guys. I am using the model trained using mmdet:v3.3.0 and using mmdeploy:v1.3.1, onnxruntime:1.16.3. I found that the model trained using co-detr official repo (mmdet-2.25) requires lot of tinkering as the model backbone(swintransformer-v1 vs v2) is little different and also preprocessing functions and some utilities functions required for inference are different. So, if you train a new model from mmdet:v3.3.0 repo, I think you will be able to export it to onnx.
To solve this issue:
RuntimeError:Only tuples, lists and Variables are supported as JIT inputs/outputs. Dictionaries and strings are also accepted, but their usage is not recommended. Here, received an input of unsupported type: DetDataSample
I modified ...(your-env-path)/site-packages/torch/jit/trace.py
In line 125, I added try except as a temporary fix. I am using the tuple format so that the code doesn't break later during visualization. I am following the same format as the plain dino export to onnx. The change is like this:
try:
result_ = self.inner(*trace_inputs)
result = result_[0]
scores, bboxes, labels = result.pred_instances.scores, result.pred_instances.bboxes, result.pred_instances.labels
# Combining scores and bboxes into the dino format
combined_tensor = torch.cat((bboxes, scores.unsqueeze(1)), dim=1)
formatted_labels = labels.unsqueeze(0).long() # Retaining integer format
# Creating tuples with combined tensors and labels
formatted_result = (combined_tensor.unsqueeze(0), formatted_labels)
outs.append(formatted_result)
except:
print(self.inner(*trace_inputs))
outs.append(self.inner(*trace_inputs))
Also, we need to remove soft_nms operations, just comment this section out like below in your config file:
...
test_cfg=[
dict(
max_per_img=300,
# nms=dict(iou_threshold=0.8, type='soft_nms')
),
...
Hope this helps! Thank you! :)
Hi! I figured out to convert co-dino to onnx and would like to share with you guys. I am using the model trained using mmdet:v3.3.0 and using mmdeploy:v1.3.1, onnxruntime:1.16.3. I found that the model trained using co-detr official repo (mmdet-2.25) requires lot of tinkering as the model backbone(swintransformer-v1 vs v2) is little different and also preprocessing functions and some utilities functions required for inference are different. So, if you train a new model from mmdet:v3.3.0 repo, I think you will be able to export it to onnx.
To solve this issue:
RuntimeError:Only tuples, lists and Variables are supported as JIT inputs/outputs. Dictionaries and strings are also accepted, but their usage is not recommended. Here, received an input of unsupported type: DetDataSampleI modified ...(your-env-path)/site-packages/torch/jit/trace.py In line 125, I added
try exceptas a temporary fix. I am using the tuple format so that the code doesn't break later during visualization. I am following the same format as the plain dino export to onnx. The change is like this:try: result_ = self.inner(*trace_inputs) result = result_[0] scores, bboxes, labels = result.pred_instances.scores, result.pred_instances.bboxes, result.pred_instances.labels # Combining scores and bboxes into the dino format combined_tensor = torch.cat((bboxes, scores.unsqueeze(1)), dim=1) formatted_labels = labels.unsqueeze(0).long() # Retaining integer format # Creating tuples with combined tensors and labels formatted_result = (combined_tensor.unsqueeze(0), formatted_labels) outs.append(formatted_result) except: print(self.inner(*trace_inputs)) outs.append(self.inner(*trace_inputs))Also, we need to remove soft_nms operations, just comment this section out like below in your config file:
... test_cfg=[ dict( max_per_img=300, # nms=dict(iou_threshold=0.8, type='soft_nms') ), ...Hope this helps! Thank you! :)
cool, I will try
try: result_ = self.inner(*trace_inputs) result = result_[0] scores, bboxes, labels = result.pred_instances.scores, result.pred_instances.bboxes, result.pred_instances.labels # Combining scores and bboxes into the dino format combined_tensor = torch.cat((bboxes, scores.unsqueeze(1)), dim=1) formatted_labels = labels.unsqueeze(0).long() # Retaining integer format # Creating tuples with combined tensors and labels formatted_result = (combined_tensor.unsqueeze(0), formatted_labels) outs.append(formatted_result) except: print(self.inner(*trace_inputs)) outs.append(self.inner(*trace_inputs))
is pytoch2torchscript can modified ...(your-env-path)/site-packages/torch/jit/trace.py to slove `RuntimeError: Tracer cannot infer type of [<DetDataSample(
META INFORMATION
scale_factor: (2.9984375, 2.997658079625293)
img_id: 0
ori_shape: (427, 640)
batch_input_shape: (1280, 1919)
img_shape: (1280, 1919)
img_path: '/mnt/data/train-yolov7-images/pytorch2onn-mmdeploy/mmdetection-support_dino_onnx/mmdetection-support_dino_onnx/demo/demo.jpg'
pad_shape: (1280, 1919)
DATA FIELDS
ignored_instances: <InstanceData(
META INFORMATION
DATA FIELDS
) at 0x7f2ef4d998b0>
pred_instances: <InstanceData(
META INFORMATION
DATA FIELDS
labels: tensor([ 0, 0, 0, 0, 0, 25, 0, 0, 0, 0, 16, 2, 2, 25, 2, 25, 25, 0,
16, 29, 0, 14, 2, 0, 2, 16, 0, 0, 16, 16, 10, 25, 25, 14, 25, 2,
0, 25, 2, 2, 37, 58, 25, 14, 10, 14, 0, 2, 58, 26, 10, 25, 2, 25,
2, 2, 58, 25, 10, 25, 58, 3, 2, 33, 0, 25, 58, 58, 0, 0, 0, 9,
0, 56, 26, 14, 0, 25, 0, 2, 25, 2, 17, 25, 25, 25, 25, 11, 25, 0,
0, 58, 0, 2, 14, 58, 0, 25, 25, 29, 2, 25, 2, 58, 2, 40, 2, 3,
26, 3, 9, 38, 25, 9, 25, 3, 33, 0, 0, 17, 55, 11, 10, 58, 58, 16,
16, 16, 2, 56, 36, 26, 58, 17, 17, 56, 0, 35, 0, 12, 12, 25, 2, 2,
25, 77, 8, 9, 46, 36, 4, 36, 25, 2, 26, 2, 9, 0, 0, 15, 25, 0,
25, 2, 14, 33, 0, 0, 10, 8, 60, 33, 25, 58, 13, 26, 0, 0, 0, 0,
0, 0, 25, 2, 25, 2, 58, 0, 2, 25, 25, 14, 2, 14, 16, 25, 25, 26,
0, 0, 0, 2, 0, 9, 58, 3, 26, 25, 17, 58, 2, 2, 56, 25, 14, 2,
0, 25, 16, 10, 25, 2, 58, 16, 25, 25, 58, 25, 2, 0, 0, 2, 0, 25,
2, 0, 0])
scores: tensor([0.2160, 0.1991, 0.1698, 0.1452, 0.1358, 0.1283, 0.0971, 0.0791, 0.0758,
0.0723, 0.0697, 0.0694, 0.0666, 0.0625, 0.0615, 0.0583, 0.0566, 0.0557,
0.0551, 0.0535, 0.0527, 0.0525, 0.0512, 0.0511, 0.0509, 0.0474, 0.0474,
0.0470, 0.0468, 0.0457, 0.0454, 0.0446, 0.0443, 0.0439, 0.0435, 0.0432,
0.0431, 0.0429, 0.0428, 0.0419, 0.0400, 0.0399, 0.0398, 0.0397, 0.0396,
0.0393, 0.0390, 0.0386, 0.0383, 0.0379, 0.0377, 0.0376, 0.0373, 0.0369,
0.0359, 0.0354, 0.0351, 0.0351, 0.0347, 0.0342, 0.0341, 0.0340, 0.0337,
0.0336, 0.0335, 0.0334, 0.0333, 0.0331, 0.0331, 0.0323, 0.0323, 0.0320,
0.0317, 0.0312, 0.0312, 0.0310, 0.0309, 0.0305, 0.0302, 0.0301, 0.0300,
0.0299, 0.0299, 0.0297, 0.0297, 0.0296, 0.0294, 0.0292, 0.0292, 0.0291,
0.0290, 0.0290, 0.0289, 0.0287, 0.0287, 0.0286, 0.0286, 0.0285, 0.0283,
0.0282, 0.0282, 0.0282, 0.0281, 0.0281, 0.0280, 0.0278, 0.0276, 0.0276,
0.0276, 0.0275, 0.0271, 0.0270, 0.0269, 0.0269, 0.0269, 0.0268, 0.0267,
0.0266, 0.0266, 0.0263, 0.0263, 0.0262, 0.0258, 0.0258, 0.0257, 0.0256,
0.0255, 0.0254, 0.0253, 0.0249, 0.0249, 0.0248, 0.0247, 0.0246, 0.0244,
0.0243, 0.0242, 0.0242, 0.0241, 0.0240, 0.0240, 0.0239, 0.0239, 0.0237,
0.0237, 0.0235, 0.0235, 0.0235, 0.0235, 0.0232, 0.0232, 0.0229, 0.0229,
0.0229, 0.0228, 0.0228, 0.0227, 0.0227, 0.0226, 0.0224, 0.0224, 0.0223,
0.0222, 0.0222, 0.0222, 0.0221, 0.0220, 0.0220, 0.0219, 0.0219, 0.0219,
0.0218, 0.0218, 0.0218, 0.0218, 0.0217, 0.0217, 0.0194, 0.0167, 0.0114,
0.0099, 0.0099, 0.0088, 0.0084, 0.0072, 0.0069, 0.0064, 0.0060, 0.0056,
0.0055, 0.0052, 0.0052, 0.0051, 0.0049, 0.0049, 0.0048, 0.0044, 0.0043,
0.0043, 0.0042, 0.0042, 0.0041, 0.0039, 0.0039, 0.0038, 0.0037, 0.0037,
0.0037, 0.0036, 0.0035, 0.0035, 0.0034, 0.0034, 0.0032, 0.0032, 0.0031,
0.0031, 0.0030, 0.0029, 0.0029, 0.0028, 0.0028, 0.0028, 0.0026, 0.0025,
0.0025, 0.0023, 0.0023, 0.0023, 0.0021, 0.0020, 0.0016, 0.0016, 0.0015,
0.0013, 0.0011, 0.0010])
bboxes: tensor([[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[0.0000e+00, 7.2158e+01, 3.8210e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 6.0318e+01, 2.8487e+02, 4.2700e+02],
[2.6185e+02, 7.1847e+01, 6.4000e+02, 4.2432e+02],
[2.9458e+02, 1.4595e+02, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 5.4273e+01, 2.9559e+02, 3.2202e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.4563e+02, 1.4659e+01, 6.3662e+02, 3.9578e+02],
[4.7329e+00, 4.1630e+01, 6.3977e+02, 3.9011e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[2.1382e+02, 1.6523e+02, 6.4000e+02, 4.2700e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[2.1120e+02, 1.2138e+02, 6.3890e+02, 3.8925e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[2.4776e+02, 1.5199e+01, 6.4000e+02, 3.9079e+02],
[0.0000e+00, 4.1204e+01, 3.8788e+02, 3.6426e+02],
[4.6684e+02, 7.1337e+01, 6.4000e+02, 4.2491e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 1.6564e+02, 5.1611e+02, 4.2700e+02],
[0.0000e+00, 1.5713e+02, 3.6014e+02, 4.2700e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 8.2242e-01, 3.9962e+02, 3.2408e+02],
[0.0000e+00, 5.4273e+01, 2.9559e+02, 3.2202e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[3.2242e+02, 7.0608e+01, 6.4000e+02, 2.5587e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[1.0147e+01, 1.5368e+02, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 4.1166e+01, 4.4898e+02, 3.7798e+02],
[5.6229e+02, 4.6781e+01, 6.4000e+02, 4.2581e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 6.9887e+00, 2.8779e+02, 4.1326e+02],
[0.0000e+00, 3.9866e+01, 2.5395e+02, 3.0416e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 5.4554e+01, 5.3191e+02, 4.1821e+02],
[0.0000e+00, 6.9887e+00, 2.8779e+02, 4.1326e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[1.6044e+02, 5.3330e+01, 5.0755e+02, 3.8331e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[2.5501e+01, 6.2984e+00, 6.4000e+02, 3.8557e+02],
[0.0000e+00, 8.2242e-01, 3.9962e+02, 3.2408e+02],
[0.0000e+00, 5.4273e+01, 2.9559e+02, 3.2202e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 5.4554e+01, 5.3191e+02, 4.1821e+02],
[0.0000e+00, 7.2158e+01, 3.8210e+02, 4.2700e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[2.1382e+02, 1.6523e+02, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 1.3510e+02, 3.0765e+02, 4.1102e+02],
[1.5868e+02, 2.7586e+01, 6.4000e+02, 3.8983e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[4.2014e+02, 1.6466e+02, 6.4000e+02, 3.7918e+02],
[3.8734e+02, 1.7227e+02, 6.4000e+02, 4.2472e+02],
[0.0000e+00, 3.9866e+01, 2.5395e+02, 3.0416e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[4.4463e-01, 0.0000e+00, 1.0380e+02, 4.2604e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[1.6612e+00, 1.8108e+02, 1.3982e+02, 4.2700e+02],
[0.0000e+00, 1.2050e+00, 3.0532e+02, 2.0350e+02],
[0.0000e+00, 1.5517e+02, 2.3322e+02, 4.2700e+02],
[3.8747e+00, 0.0000e+00, 6.4000e+02, 2.2986e+02],
[1.6039e+02, 7.8185e+01, 4.4346e+02, 2.8939e+02],
[0.0000e+00, 5.4273e+01, 2.9559e+02, 3.2202e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 6.9887e+00, 2.8779e+02, 4.1326e+02],
[7.7882e+01, 1.5648e+01, 4.9433e+02, 3.7667e+02],
[3.8785e+02, 8.7845e+01, 5.7135e+02, 2.0960e+02],
[0.0000e+00, 6.2847e-01, 2.2725e+02, 1.3132e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 0.0000e+00, 5.2475e+02, 1.3465e+02],
[2.0440e+02, 3.4097e+01, 6.4000e+02, 3.8894e+02],
[5.7069e+02, 1.0939e+02, 5.8223e+02, 1.1748e+02],
[0.0000e+00, 4.1204e+01, 3.8788e+02, 3.6426e+02],
[0.0000e+00, 8.3078e+00, 2.6751e+02, 3.6747e+02],
[1.5868e+02, 2.7586e+01, 6.4000e+02, 3.8983e+02],
[2.6185e+02, 7.1847e+01, 6.4000e+02, 4.2432e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[3.5561e+00, 0.0000e+00, 6.4000e+02, 2.4098e+02],
[1.0667e+02, 6.9062e+01, 4.2385e+02, 3.6903e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[6.2002e+01, 0.0000e+00, 6.2688e+02, 1.4552e+02],
[4.1718e+00, 4.1516e-01, 1.7336e+02, 1.3076e+02],
[0.0000e+00, 0.0000e+00, 5.5502e+02, 1.4696e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[4.4757e+02, 2.5681e+01, 6.4000e+02, 4.2443e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[3.3409e+02, 6.9283e+01, 5.2455e+02, 3.9823e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[2.3454e+02, 8.3355e+01, 6.4000e+02, 3.8414e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[1.7801e+02, 2.0820e+02, 6.4000e+02, 4.2700e+02],
[4.0794e+02, 5.2874e+01, 6.4000e+02, 3.9537e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[4.4463e-01, 0.0000e+00, 1.0380e+02, 4.2604e+02],
[0.0000e+00, 8.7248e+01, 2.9231e+02, 4.2453e+02],
[2.6185e+02, 7.1847e+01, 6.4000e+02, 4.2432e+02],
[5.7069e+02, 1.0939e+02, 5.8223e+02, 1.1748e+02],
[2.1382e+02, 1.6523e+02, 6.4000e+02, 4.2700e+02],
[5.8778e+02, 1.0198e+01, 6.4000e+02, 4.0520e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[0.0000e+00, 3.9866e+01, 2.5395e+02, 3.0416e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[5.8725e+02, 8.0794e+01, 5.9527e+02, 8.9090e+01],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[5.6229e+02, 4.6781e+01, 6.4000e+02, 4.2581e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.9076e+02, 8.4498e+01, 6.3985e+02, 4.1969e+02],
[4.4463e-01, 0.0000e+00, 1.0380e+02, 4.2604e+02],
[0.0000e+00, 7.2158e+01, 3.8210e+02, 4.2700e+02],
[2.9458e+02, 1.4595e+02, 6.4000e+02, 4.2700e+02],
[5.8774e+01, 2.0932e+00, 2.9747e+02, 3.0555e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[7.4486e+01, 1.8788e+01, 3.0107e+02, 3.5681e+02],
[0.0000e+00, 4.1204e+01, 3.8788e+02, 3.6426e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[4.2666e+02, 1.7290e+02, 6.4000e+02, 4.2700e+02],
[5.6229e+02, 4.6781e+01, 6.4000e+02, 4.2581e+02],
[0.0000e+00, 4.0296e+01, 4.7342e+02, 3.8573e+02],
[2.4282e+00, 2.4388e+00, 6.3059e+01, 4.2700e+02],
[1.6500e+02, 7.7726e+01, 6.4000e+02, 4.2700e+02],
[3.8403e+02, 3.8851e+00, 6.3406e+02, 1.7161e+02],
[0.0000e+00, 8.2242e-01, 3.9962e+02, 3.2408e+02],
[3.0750e+02, 7.8088e+01, 4.9292e+02, 4.0430e+02],
[2.8905e+00, 4.7422e-01, 4.8275e+02, 9.4585e+01],
[5.7069e+02, 1.0939e+02, 5.8223e+02, 1.1748e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[5.8047e+02, 8.1446e+01, 5.9600e+02, 9.0194e+01],
[0.0000e+00, 2.8271e+02, 4.8164e+02, 4.2694e+02],
[0.0000e+00, 6.0318e+01, 2.8487e+02, 4.2700e+02],
[0.0000e+00, 7.3228e+01, 2.9850e+02, 2.0027e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[2.4165e+02, 1.6792e+01, 6.4000e+02, 3.9862e+02],
[0.0000e+00, 8.2242e-01, 3.9962e+02, 3.2408e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[0.0000e+00, 5.4273e+01, 2.9559e+02, 3.2202e+02],
[0.0000e+00, 1.1854e+02, 1.3432e+02, 3.3936e+02],
[4.7329e+00, 4.1630e+01, 6.3977e+02, 3.9011e+02],
[3.9284e+02, 9.1945e+01, 6.4000e+02, 3.8028e+02],
[0.0000e+00, 7.4748e+01, 4.2332e+02, 4.2700e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[0.0000e+00, 6.9589e+01, 2.9147e+02, 3.7202e+02],
[1.0793e+00, 7.7588e+01, 6.4000e+02, 4.2700e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[0.0000e+00, 4.1204e+01, 3.8788e+02, 3.6426e+02],
[2.8507e+02, 1.1117e+01, 6.4000e+02, 3.7584e+02],
[4.7329e+00, 4.1630e+01, 6.3977e+02, 3.9011e+02],
[0.0000e+00, 1.8234e+02, 4.6567e+02, 4.2700e+02],
[3.9010e+02, 1.0139e+01, 6.4000e+02, 4.0455e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[0.0000e+00, 8.8411e-01, 2.8525e+02, 3.2094e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[2.6185e+02, 7.1847e+01, 6.4000e+02, 4.2432e+02],
[4.7329e+00, 4.1630e+01, 6.3977e+02, 3.9011e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[0.0000e+00, 2.4596e+00, 3.6105e+02, 2.9753e+02],
[1.5847e+02, 6.8899e+01, 4.7434e+02, 3.8014e+02],
[4.7329e+00, 4.1630e+01, 6.3977e+02, 3.9011e+02],
[4.3726e+02, 1.6330e+02, 6.4000e+02, 3.6836e+02],
[4.1968e+02, 9.7537e+01, 6.4000e+02, 4.2626e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[0.0000e+00, 2.4596e+00, 3.6105e+02, 2.9753e+02],
[4.2666e+02, 1.7290e+02, 6.4000e+02, 4.2700e+02],
[3.3403e+02, 7.9988e+01, 6.4000e+02, 4.2429e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[2.5501e+01, 6.2984e+00, 6.4000e+02, 3.8557e+02],
[3.2040e+02, 6.6847e+01, 6.4000e+02, 3.8979e+02],
[1.6356e+00, 5.5374e+01, 3.0322e+02, 3.5976e+02],
[5.5122e+02, 4.9658e+01, 6.4000e+02, 4.1820e+02],
[1.0730e+02, 0.0000e+00, 6.3086e+02, 1.3961e+02],
[3.3831e+02, 7.2650e+01, 6.4000e+02, 4.2263e+02],
[3.9284e+02, 9.1945e+01, 6.4000e+02, 3.8028e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[0.0000e+00, 0.0000e+00, 5.2475e+02, 1.3465e+02],
[1.4443e+02, 7.2818e+01, 6.4000e+02, 4.2557e+02],
[0.0000e+00, 0.0000e+00, 5.5502e+02, 1.4696e+02],
[2.8250e+02, 4.3829e+01, 6.3907e+02, 4.2441e+02],
[2.1494e+02, 7.4293e+01, 6.4000e+02, 4.2397e+02],
[2.4880e-01, 7.6363e-01, 2.4526e+02, 1.3574e+02],
[5.8323e+02, 9.6534e+00, 6.4000e+02, 4.2157e+02],
[2.1494e+02, 7.4293e+01, 6.4000e+02, 4.2397e+02],
[2.5619e+02, 1.6856e+02, 6.4000e+02, 4.2599e+02],
[0.0000e+00, 4.0296e+01, 4.7342e+02, 3.8573e+02],
[1.7174e+02, 8.9912e+01, 4.4728e+02, 2.8728e+02],
[0.0000e+00, 7.4748e+01, 4.2332e+02, 4.2700e+02],
[4.6382e+00, 0.0000e+00, 6.4000e+02, 2.1783e+02],
[0.0000e+00, 7.4748e+01, 4.2332e+02, 4.2700e+02],
[2.3454e+02, 8.3355e+01, 6.4000e+02, 3.8414e+02],
[0.0000e+00, 7.5742e+01, 3.5057e+02, 4.2451e+02],
[1.4443e+02, 7.2818e+01, 6.4000e+02, 4.2557e+02],
[0.0000e+00, 6.9887e+00, 2.8779e+02, 4.1326e+02],
[1.4443e+02, 7.2818e+01, 6.4000e+02, 4.2557e+02],
[4.6382e+00, 0.0000e+00, 6.4000e+02, 2.1783e+02],
[0.0000e+00, 8.8411e-01, 2.8525e+02, 3.2094e+02],
[1.6356e+00, 5.5374e+01, 3.0322e+02, 3.5976e+02]])
) at 0x7f2e5428f8e0>
gt_instances: <InstanceData(
META INFORMATION
DATA FIELDS
) at 0x7f2ef4d99880>
) at 0x7f2ef4d99970>] :Could not infer type of list element: Only tensors and (possibly nested) tuples of tensors, lists, or dictsare supported as inputs or outputs of traced functions, but instead got value of type DetDataSample. `
@xinlin-xiao Mm it seems it is returning in list format, ao just get the entry at 0th index of that list and it should solve your issue
Hi @bibekyess, could you please provide the entire _trace.py file? mine seems to look different from yours, as I can't add anything to line 125 (that line for me is in the middle of a function call).
Moreover, I'm using the implementation that's included in mmdetection v3.3.0 and I can confirm that issue is present also here.
Hi @sirolf-otrebla I am now on vacation so I will provide it later. But as far as I remember, line 125 includes this instruction outs.append(self.inner(*trace_inputs)) Now just substitute that with the above try-except snippet. Or you can just search for that instruction in this file and substitute.