Jackson Ding
Jackson Ding
> use AdamW with smaller learning rate > > see finetuning section in readme Thank you!
> Hi - to apply the above is this batch_size and max_len ? Noting, I was unable to fix OOM on small GPUs using the prescribed methods: export 'PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512' (fill...
> @WoBuChiTang hi, i need to train this model on a long audio dataset (up to 20 second long), curious what's the max `max_num_tokens` you were able to pull off...
> Hi there. I saw that the repo's code only support Engilish aligner training expand vocab size that including all of chinese and english phonemes, prepare dataset as train.txt
> Yes we will make the model support Chinese in the future I'm looking forward to this. Thank you for this great work.