quick_sentence_transformers icon indicating copy to clipboard operation
quick_sentence_transformers copied to clipboard

Published 20 hours ago • verified publisher icon yuanzhoulvpi2017

推理时模型维度报错

Open ysujiang opened this issue 2 years ago • 3 comments

File "torch_to_onnx.py", line 190, in conver_bert_torch_to_onnx() File "torch_to_onnx.py", line 104, in conver_bert_torch_to_onnx sess = onnxruntime.InferenceSession(MODEL_ONNX_PATH) File "/opt/conda/lib/python3.7/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 335, in init self._create_inference_session(providers, provider_options, disabled_optimizers) File "/opt/conda/lib/python3.7/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 379, in _create_inference_session sess.initialize_session(providers, provider_options, disabled_optimizers) onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Node (Squeeze_1169) Op (Squeeze) [ShapeInferenceError] Dimension of input 1 must be 1 instead of 768 有遇到过这样的问题吗

ysujiang avatar Mar 10 '22 07:03 ysujiang

是不是模型选的不对，这个只是针对那一个模型做了加速，如果换别的模型，需要修改代码，目前我还没完备代码

yuanzhoulvpi2017 avatar Mar 10 '22 07:03 yuanzhoulvpi2017

我是用的pytorch_model.bin这个模型，需要改那一部分代码能否告知一下

ysujiang avatar Mar 10 '22 08:03 ysujiang

哈哈，只是针对某一个模型才有用的~

zlszhonglongshen avatar Sep 01 '22 07:09 zlszhonglongshen