[Bug] Yoloxpose模型导出为 TensorRT 格式错误 #2500
Comments
|
还有一个小问题，在此之前，我已成功将rtmpose部署导出为TensorRT 格式，且成功测试推理 （所以应该不是环境设置的问题吧），Yoloxpose部署导出onnx格式成功，但只能使用API推理测试，无法使用SDK推理。 [2023-10-17 15:59:07.788] [mmdeploy] [info] [model.cpp:35] [DirectoryModel] Load model: "mmdeploy-model/yoloxpose-ort/" |
|
you need to rebuild mmdeploy. |
不是通过 pip install -e {dir}/mmdeploy 重新构建嘛 |
|
Yes. you have build from source: https://mmdeploy.readthedocs.io/en/latest/01-how-to-build/build_from_source.html |
|
This issue is marked as stale because it has been marked as invalid or awaiting response for 7 days without any further response. It will be closed in 5 days if the stale label is not removed or if there is no further response. |
|
This issue is closed because it has been stale for 5 days. Please open a new issue if you have similar issues or you have any new updates now. |
Checklist
Describe the bug
10/17 15:23:31 - mmengine - WARNING - Failed to search registry with scope "mmpose" in the "Codebases" registry tree. As a workaround, the current "Codebases" registry in "mmdeploy" is used to build instance. This may cause unexpected failure when running the built modules. Please check whether "mmpose" is a correct scope, or whether the registry is initialized.
10/17 15:23:31 - mmengine - WARNING - Failed to search registry with scope "mmpose" in the "mmpose_tasks" registry tree. As a workaround, the current "mmpose_tasks" registry in "mmdeploy" is used to build instance. This may cause unexpected failure when running the built modules. Please check whether "mmpose" is a correct scope, or whether the registry is initialized.
10/17 15:23:33 - mmengine - INFO - Start pipeline mmdeploy.apis.pytorch2onnx.torch2onnx in subprocess
10/17 15:23:34 - mmengine - WARNING - Failed to search registry with scope "mmpose" in the "Codebases" registry tree. As a workaround, the current "Codebases" registry in "mmdeploy" is used to build instance. This may cause unexpected failure when running the built modules. Please check whether "mmpose" is a correct scope, or whether the registry is initialized.
10/17 15:23:34 - mmengine - WARNING - Failed to search registry with scope "mmpose" in the "mmpose_tasks" registry tree. As a workaround, the current "mmpose_tasks" registry in "mmdeploy" is used to build instance. This may cause unexpected failure when running the built modules. Please check whether "mmpose" is a correct scope, or whether the registry is initialized.
/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/site-packages/mmpose/datasets/datasets/utils.py:102: UserWarning: The metainfo config file "configs/base/datasets/coco.py" does not exist. A matched config file "/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/site-packages/mmpose/.mim/configs/base/datasets/coco.py" will be used instead.
warnings.warn(
Loads checkpoint by http backend from path: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-84e9a538_20230829.pth
10/17 15:23:42 - mmengine - WARNING - DeprecationWarning: get_onnx_config will be deprecated in the future.
10/17 15:23:42 - mmengine - INFO - Export PyTorch model to ONNX: mmdeploy-model/yoloxpose-trt/end2end.onnx.
10/17 15:23:43 - mmengine - WARNING - Can not find torch.nn.functional.scaled_dot_product_attention, function rewrite will not be applied
10/17 15:23:43 - mmengine - WARNING - Can not find torch._C._jit_pass_onnx_autograd_function_process, function rewrite will not be applied
10/17 15:23:43 - mmengine - WARNING - Can not find torch._C._jit_pass_onnx_deduplicate_initializers, function rewrite will not be applied
10/17 15:23:43 - mmengine - WARNING - Can not find mmdet.models.utils.transformer.PatchMerging.forward, function rewrite will not be applied
/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/site-packages/torch/functional.py:445: 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:2157.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
/home/ccnu-train/hwk/mmdeploy/mmdeploy/core/optimizers/function_marker.py:160: 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!
ys_shape = tuple(int(s) for s in ys.shape)
/home/ccnu-train/hwk/mmdeploy/mmdeploy/mmcv/ops/nms.py:475: 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!
int(scores.shape[-1]),
/home/ccnu-train/hwk/mmdeploy/mmdeploy/mmcv/ops/nms.py:149: 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!
out_boxes = min(num_boxes, after_topk)
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/site-packages/torch/onnx/symbolic_opset9.py:2815: UserWarning: Exporting aten::index operator of advanced indexing in opset 11 is achieved by combination of multiple ONNX operators, including Reshape, Transpose, Concat, and Gather. If indices include negative values, the exported graph will produce incorrect results.
warnings.warn("Exporting aten::index operator of advanced indexing in opset " +
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
WARNING: The shape inference of mmdeploy::TRTBatchedNMS type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
10/17 15:23:58 - mmengine - INFO - Execute onnx optimize passes.
10/17 15:23:58 - mmengine - INFO - Finish pipeline mmdeploy.apis.pytorch2onnx.torch2onnx
10/17 15:24:01 - mmengine - INFO - Start pipeline mmdeploy.apis.utils.utils.to_backend in subprocess
10/17 15:24:01 - mmengine - WARNING - Could not load the library of tensorrt plugins. Because the file does not exist:
[10/17/2023-15:24:02] [TRT] [I] [MemUsageChange] Init CUDA: CPU +14, GPU +0, now: CPU 101, GPU 18991 (MiB)
[10/17/2023-15:24:05] [TRT] [I] [MemUsageChange] Init builder kernel library: CPU +546, GPU +118, now: CPU 702, GPU 19109 (MiB)
[10/17/2023-15:24:05] [TRT] [I] ----------------------------------------------------------------
[10/17/2023-15:24:05] [TRT] [I] Input filename: mmdeploy-model/yoloxpose-trt/end2end.onnx
[10/17/2023-15:24:05] [TRT] [I] ONNX IR version: 0.0.7
[10/17/2023-15:24:05] [TRT] [I] Opset version: 11
[10/17/2023-15:24:05] [TRT] [I] Producer name: pytorch
[10/17/2023-15:24:05] [TRT] [I] Producer version: 1.10
[10/17/2023-15:24:05] [TRT] [I] Domain:
[10/17/2023-15:24:05] [TRT] [I] Model version: 0
[10/17/2023-15:24:05] [TRT] [I] Doc string:
[10/17/2023-15:24:05] [TRT] [I] ----------------------------------------------------------------
[10/17/2023-15:24:05] [TRT] [W] onnx2trt_utils.cpp:377: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[10/17/2023-15:24:05] [TRT] [W] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[10/17/2023-15:24:06] [TRT] [I] No importer registered for op: TRTBatchedNMS. Attempting to import as plugin.
[10/17/2023-15:24:06] [TRT] [I] Searching for plugin: TRTBatchedNMS, plugin_version: 1, plugin_namespace:
[10/17/2023-15:24:06] [TRT] [E] ModelImporter.cpp:726: While parsing node number 656 [TRTBatchedNMS -> "1741"]:
[10/17/2023-15:24:06] [TRT] [E] ModelImporter.cpp:727: --- Begin node ---
[10/17/2023-15:24:06] [TRT] [E] ModelImporter.cpp:728: input: "1740"
input: "1722"
output: "1741"
output: "1742"
output: "1743"
name: "TRTBatchedNMS_656"
op_type: "TRTBatchedNMS"
attribute {
name: "background_label_id"
i: -1
type: INT
}
attribute {
name: "clip_boxes"
i: 0
type: INT
}
attribute {
name: "iou_threshold"
f: 0.65
type: FLOAT
}
attribute {
name: "is_normalized"
i: 0
type: INT
}
attribute {
name: "keep_topk"
i: 100
type: INT
}
attribute {
name: "num_classes"
i: 1
type: INT
}
attribute {
name: "return_index"
i: 1
type: INT
}
attribute {
name: "score_threshold"
f: 0.5
type: FLOAT
}
attribute {
name: "topk"
i: 5000
type: INT
}
domain: "mmdeploy"
[10/17/2023-15:24:06] [TRT] [E] ModelImporter.cpp:729: --- End node ---
[10/17/2023-15:24:06] [TRT] [E] ModelImporter.cpp:732: ERROR: builtin_op_importers.cpp:5428 In function importFallbackPluginImporter:
[8] Assertion failed: creator && "Plugin not found, are the plugin name, version, and namespace correct?"
Process Process-3:
Traceback (most recent call last):
File "/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
self.run()
File "/home/ccnu-train/anaconda3/envs/mmdeploy/lib/python3.8/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/home/ccnu-train/hwk/mmdeploy/mmdeploy/apis/core/pipeline_manager.py", line 107, in call
ret = func(*args, **kwargs)
File "/home/ccnu-train/hwk/mmdeploy/mmdeploy/apis/utils/utils.py", line 98, in to_backend
return backend_mgr.to_backend(
File "/home/ccnu-train/hwk/mmdeploy/mmdeploy/backend/tensorrt/backend_manager.py", line 127, in to_backend
onnx2tensorrt(
File "/home/ccnu-train/hwk/mmdeploy/mmdeploy/backend/tensorrt/onnx2tensorrt.py", line 79, in onnx2tensorrt
from_onnx(
File "/home/ccnu-train/hwk/mmdeploy/mmdeploy/backend/tensorrt/utils.py", line 185, in from_onnx
raise RuntimeError(f'Failed to parse onnx, {error_msgs}')
RuntimeError: Failed to parse onnx, In node 656 (importFallbackPluginImporter): UNSUPPORTED_NODE: Assertion failed: creator && "Plugin not found, are the plugin name, version, and namespace correct?"
Reproduction
python tools/deploy.py
configs/mmpose/pose-detection_yolox-pose_tensorrt_dynamic-640x640.py
../mmpose/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_m_8xb32-300e_coco-640.py
https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-84e9a538_20230829.pth
demo/resources/human-pose.jpg
--work-dir mmdeploy-model/yoloxpose-trt
--device cuda
--show
--dump-info
Environment
Error traceback
No response
The text was updated successfully, but these errors were encountered: