Aaqib

@mossaab0 If you can provide some reproduction steps, I can try to rootcause this

@yayuntian Can you provide the reproduction steps, particularly regarding the model ` --serialized-file models/bert_model.pt --extra-files ./bert_record.py,./models/content_id_map.json`. I want to get a breakdown of where the latency is added, is it...

Closing in favour of #1954

@ridhwan-saal This typically happens when there is an error in the handler file. It is possible that the logs are not getting flushed to the frontend before the worker dies....

@ridhwan-saal Thanks for the info. I will try and reproduce this. Can you also let me know what name or path you are loading here `self.config = AutoConfig.from_pretrained(model_path)`?

@ridhwan-saal I am not able to reproduce this. Let me know if I got something wrong from your instructions `torchserve --start --model-store model_store --ts-config config.properties --ncs --models all` Works fine....

@Shin-ichi-Takayama Looking at the logs, it seems like you have not installed espnet2. `ModuleNotFoundError: No module named 'espnet2'`

@Shin-ichi-Takayama Can you also confirm if the python executable being used by Torchserve `/usr/bin/python3.8` is the same as the on you are checking the `pip list` for?

From the logs it seems like it is missing a file to load ``` 2022-08-13T13:30:14,703 [INFO ] W-9000-denoise_transformer_1.0-stdout MODEL_LOG - RuntimeError: PytorchStreamReader failed locating file constants.pkl: file not found ```

@ArgoHA Can you provide the exact command you used to create the mar file?