load_model_ensemble() got an unexpected keyword argument 'model_state'??
See original GitHub issueWhen trying to evaluate pretrained models I keep getting the below error. The same error occurs for the multi-lingual model and the wav2vec_small_100h.pt model… I didn’t finetune the models as I wanted to check the fairseq provided models first. I’m using the same dictionary and using the fairseq transformer LM. Flashlight python bindings are install correctly and I built fairseq from source using pip install --editable ./ from the master directory.
python ../speech_recognition/infer.py ./english_test/text/ --task audio_pretraining --nbest 1 --path ./pretrained/xlsr_53_56k.pt --results-path ./english_test/ --w2l-decoder fairseqlm --lm-model ./pretrained/lm_librispeech_word_transformer.pt --lm-weight 2 --word-score -1 --sil-weight 0 --criterion ctc --labels ltr
2021-01-11 00:47:43 | INFO | main | | decoding with criterion ctc 2021-01-11 00:47:43 | INFO | main | | loading model(s) from ./pretrained/xlsr_53_56k.pt Traceback (most recent call last): File “…/speech_recognition/infer.py”, line 428, in <module> cli_main() File “…/speech_recognition/infer.py”, line 424, in cli_main main(args) File “…/speech_recognition/infer.py”, line 230, in main models, saved_cfg = checkpoint_utils.load_model_ensemble( TypeError: load_model_ensemble() got an unexpected keyword argument ‘model_state’
It seems like a problem with loading the model, but I can’t figure out what I did wrong?
Issue Analytics
- State:
- Created 3 years ago
- Reactions:1
- Comments:6
Top GitHub Comments
Hi, I have the same issue. Did anyone solve this problem?
Also having the same issue