# Onnx Errors pipeline\_name ='question-answering' **URL:** **Author:** ![NhatPham](https://avatars.discourse-cdn.com/v4/letter/n/94ad74/32.png) [@NhatPham](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FNhatPham)%5C **Post date:** [November 22, 2021, 6:37am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F1 "2021-11-22T06:37:15Z") from transformers.convert\_graph\_to\_onnx import convert convert(framework=‘pt’,pipeline\_name =‘question-answering’, model=‘roberta-base-squad2’,output=my\_outputpath,opset=11) ONNX opset version set to: 11 Loading pipeline (model: roberta-base-squad2, tokenizer: roberta-base-squad2) Using framework PyTorch: 1.10.0+cu111 Found input input\_ids with shape: {0: ‘batch’, 1: ‘sequence’} Found input attention\_mask with shape: {0: ‘batch’, 1: ‘sequence’} Found output output\_0 with shape: {0: ‘batch’, 1: ‘sequence’} Found output output\_1 with shape: {0: ‘batch’, 1: ‘sequence’} Ensuring inputs are in correct order token\_type\_ids is not present in the generated input list. Generated inputs order: [‘input\_ids’, ‘attention\_mask’] /usr/local/lib/python3.7/dist-packages/torch/onnx/utils.py:90: UserWarning: ‘enable\_onnx\_checker’ is deprecated and ignored. It will be removed in the next PyTorch release. To proceed despite ONNX checker failures, catch torch.onnx.ONNXCheckerError. warnings.warn("‘enable\_onnx\_checker’ is deprecated and ignored. It will be removed in " /usr/local/lib/python3.7/dist-packages/torch/onnx/utils.py:103: UserWarning: `use_external_data_format' is deprecated and ignored. Will be removed in next PyTorch release. The code will work as it is False if models are not larger than 2GB, Otherwise set to False because of size limits imposed by Protocol Buffers. warnings.warn("`use\_external\_data\_format’ is deprecated and ignored. Will be removed in next " Successfully converted. however when I tried to run sessions = InferenceSession(‘onnx/roberta-base-squad2.onnx’,options) _**sessions.run([‘output\_0’,‘output\_1’],[context,ques])**_ ``` 186 output_names = [output.name for output in self._outputs_meta] 187 try: ``` → 188 return self.\_sess.run(output\_names, input\_feed, run\_options) 189 except C.EPFail as err: 190 if self.\_enable\_fallback: TypeError: run(): incompatible function arguments. The following argument types are supported: 1. (self: onnxruntime.capi.onnxruntime\_pybind11\_state.InferenceSession, arg0: List[str], arg1: Dict[str, object], arg2: onnxruntime.capi.onnxruntime\_pybind11\_state.RunOptions) → List[object] Invoked with: \, [‘output\_0’, ‘output\_1’], {‘input\_ids’: tensor([[0, 2264, 16, 5, 645, 346, 116, 2]]), ‘attention\_mask’: tensor([[1, 1, 1, 1, 1, 1, 1, 1]])}, None ---
**Author:** ![Shamik](https://avatars.discourse-cdn.com/v4/letter/s/ea5d25/32.png) [@Shamik](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FShamik)%5C **Post date:** [February 23, 2022, 4:51am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F2 "2022-02-23T04:51:05Z")
@NhatPham did you manage to solve this ? i am facing the same problem as stated below ```auto --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /home/shamik/Repos/Codes/QA_ONNX.ipynb Cell 18' in ----> 1 onnx_model.run(input_feed=tokenizer(question,text), output_names=None) File ~/anaconda3/envs/onnx/lib/python3.9/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:192, in Session.run(self, output_names, input_feed, run_options) 190 output_names = [output.name for output in self._outputs_meta] 191 try: --> 192 return self._sess.run(output_names, input_feed, run_options) 193 except C.EPFail as err: 194 if self._enable_fallback: TypeError: run(): incompatible function arguments. The following argument types are supported: 1. (self: onnxruntime.capi.onnxruntime_pybind11_state.InferenceSession, arg0: List[str], arg1: Dict[str, object], arg2: onnxruntime.capi.onnxruntime_pybind11_state.RunOptions) -> List[object] Invoked with: , ['output_0', 'output_1'], {'input_ids': [101, 2040, 2001, 3958, 27227, 1029, 102, 3958, 27227, 2001, 1037, 3835, 13997, 102], 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1], 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}, None ``` ---
**Author:** ![NhatPham](https://avatars.discourse-cdn.com/v4/letter/n/94ad74/32.png) [@NhatPham](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FNhatPham)%5C **Post date:** [February 23, 2022, 8:54am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F3 "2022-02-23T08:54:45Z")
unfortunately, I decide to move on this one. 😅 ---
**Author:** ![Shamik](https://avatars.discourse-cdn.com/v4/letter/s/ea5d25/32.png) [@Shamik](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FShamik)%5C **Post date:** [February 23, 2022, 10:28am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F4 "2022-02-23T10:28:34Z")
Had some luck and managed to solve it. The `input_feed` arg while running the session for inferencing requires a dictionary object with numpy arrays and it was failing in my case because i was passing in a `BatchEncoding`(transformers.tokenization\_utils\_base.BatchEncoding) object type by doing the following: `onnx_model.run(input_feed=tokenizer(question, text), output_names=None)` **The solution is the following** ```auto inputs = tokenizer(question, text, add_special_tokens=True, return_tensors='np') # outputs = onnx_model.run(input_feed=dict(tokenizer(question, text, add_special_tokens=True, return_tensors='np')), output_names=None) outputs = onnx_model.run(input_feed=dict(inputs), output_names=None) # start_logits = onnx_model.get_outputs()[0].name # end_logits = onnx_model.get_outputs()[1].name # outputs = onnx_model.run(input_feed=dict(inputs), output_names=[start_logits,end_logits]) start_scores = outputs[0] end_scores = outputs[1] input_ids = inputs["input_ids"].tolist()[0] ans_start = np.argmax(start_scores) ans_end = np.argmax(end_scores)+1 answer_tokens = tokenizer.decode(input_ids[ans_start:ans_end]) answer_tokens ``` ---
**Author:** ![NhatPham](https://avatars.discourse-cdn.com/v4/letter/n/94ad74/32.png) [@NhatPham](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FNhatPham)%5C **Post date:** [February 23, 2022, 10:46am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F5 "2022-02-23T10:46:19Z")
haha awesounds good, I would check later Shamik and have a good day 🙂 ---
**Author:** ![NhatPham](https://avatars.discourse-cdn.com/v4/letter/n/94ad74/32.png) [@NhatPham](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FNhatPham)%5C **Post date:** [February 28, 2022, 4:18am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fonnx-errors-pipeline-name-question-answering%2F12146%2F6 "2022-02-28T04:18:53Z")
> [@Shamik](#): > > `inputs = tokenizer(question, text, add_special_tokens=True, return_tensors='np')` it works like charm thank you again Shamik