about finetune
#1
by
polarnight12
- opened
Hello,I am very interested in your work on dysarthria. I tried to fine tune the model using your parameters but the WER remains at 1. I would appreciate if you could share some of your details.
Hello, unfortunately I don't remember details today. Try this colab https://github.com/jmaczan/asr-dysarthria/blob/4311c397a499a78527a0a9cf9d592872246e5b86/notebooks/1.0-jpm-first-successful-training.ipynb
jmaczan
changed discussion status to
closed
Hello, I tried to fine-tune the model again by referring to the notebook, and the training results are as follows, it still stays at wer as 1. Have you encountered this problem? Looking forward to your reply.
