fairseq icon indicating copy to clipboard operation
fairseq copied to clipboard

hubert finetune Could not override 'task.data'.

Open liqiyao0430 opened this issue 2 years ago • 3 comments

❓ Questions and Help

Before asking:

  1. search the issues.
  2. search the docs.

What is your question?

image my code are as follows.And the path is right python fairseq_cli/hydra_train.py --config-dir C:\Users\lenovo\Desktop\github_hubert\path\to\fairseq-py\examples\hubert\config\finetune --config-name hubert_large_librispeech task.data=C:\Users\lenovo\Desktop\github_hubert\path\to\data task.label_dir=C:\Users\lenovo\Desktop\github_hubert\path\to\ltr model.w2v_path=C:\Users\lenovo\Desktop\github_hubert\path\to\the

Code

What have you tried?

What's your environment?

  • fairseq Version (e.g., 1.0 or main):
  • PyTorch Version (e.g., 1.0)
  • OS (e.g., Linux):
  • windows
  • How you installed fairseq (pip, source):
  • pip
  • Build command you used (if compiling from source):
  • Python version:
  • CUDA/cuDNN version:
  • GPU models and configuration:
  • Any other relevant information:

liqiyao0430 avatar Apr 09 '22 05:04 liqiyao0430

when i add +, it will report this problem image

liqiyao0430 avatar Apr 09 '22 05:04 liqiyao0430

--config-dir /path/to/fairseq-py/examples/hubert/config/finetune
--config-name base_10h not be change,then solve the problem

liqiyao0430 avatar Apr 09 '22 05:04 liqiyao0430

Update: It's Working now. Had an issue with relative path.

@liqiyao0430 hi there! I'm still getting the same issue, have you solved this? I used the following command:

fairseq-hydra-train distributed_training.distributed_port=2200 task.data=./wav2vec2_data model.w2v_path=pre-trained-checkpoints/checkpoint_best.pt --config-dir fairseq/examples/wav2vec/config/finetuning --config-name base_960h

shuvohishab avatar Oct 02 '22 07:10 shuvohishab