about tensor don't match size

[Ohahehe]

When I run CUDA_VISIBLE_DEVICES=0,1 python tasks/run.py --config egs/stylesinger.yaml --exp_name StyleSinger --reset to train my model based on the Chinese dataset, I got this error. I have done the work of data processing, this may be where to make config file changes. If you see this message, please help reply. Thank you very much.

[AaronZ345]

After seeing your question, I tried to replicate the steps you described. I was able to successfully complete the training process without encountering any errors. There are a few things you might want to check:

1. Data Preprocessing:
   Could you double-check that your ph_set configuration during data preprocessing is correct?
   Also, please verify that the F0 extraction process is working as intended.
2. Environment:
   Are you sure you're using the correct environment? I successfully ran the StyleSinger training in a relatively older Python environment, specifically version 3.8.19. This detail might be important.
3. TensorBoard Import Update:
   I've recently updated a tensorboard import that might have been causing the problem. Please take a look at the latest code on the repository. It could potentially resolve the issue you're facing.

Meanwhile, you can also try to use TCSinger(https://github.com/AaronZ345/TCSinger/tree/main) for style transfer trainging.