issue in predict.py while loading the trained model no config.json file

[manishhnnegi]

while loading model its not able to load model due to no config.json file present in the generated model directory.

OSError: misca does not appear to have a file named config.json. Checkout 'https://huggingface.co/misca/main' for available files.
also after loading model while prediction its asking for sequence_length and heads which is not present in the inputs dictionary.

[tmrnvcome]

i have the same issue too.

[thinhphp]

Hi, we have updated the predict.py file. You should set hyper-parameters the same as training, and have the checkpoint file (model.bin) in model_dir argument.

[tmrnvcome]

Hi. What about line 231, *if 'bert' in self.args.model_type*: This has to be changed to '*roberta*' instead right? * if 'roberta' in self.args.model_type:* Because your code uses *lstm* and *roberta* only. Regards Beatrice

…

On Wed, 27 Mar 2024, 13:28 Phu Thinh, ***@***.***> wrote: Hi, we have updated the predict.py file. You should set hyper-parameters the same as training, and have the checkpoint file (model.bin) in model_dir argument. — Reply to this email directly, view it on GitHub <#6 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BABB7ZM3S3OOHKDYCJAH2RLY2JKHPAVCNFSM6AAAAABDL5QQDOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRRHE3TCNRVHE> . You are receiving this because you commented.Message ID: ***@***.***>

[thinhphp]

Yes, you can change it to 'roberta' in this case, but our code can be used with other BERT-based PLMs.