Roya

I don't know if you've ever encountered this issue in your training, the loss increased but all other metrics decreased. It's weird... I don't know if it affects convergence.

Hello!!! I just can't replicate your reslut on snips when using bert + CRF, it's the baseline bert-base-uncased or any tricks when implement training?

I wanna to ask why loss is 0 at random in the same experimental setting when runing finetune_toy_low_resource.sh in ltu_as code or have you ever meet this problem. Do you...

Stage training sh scripts for low resource

1

I'd like to ask is there stage training sh scripts for low resource. It seems there is only low resource training sh scripts for finetune toy data. 

where to download whisper model?

2

I want to ask where can download the whisper model file end with .pt. 

train_scripts

3

I notice a lot of sh script in the directory train_scripts, I want to ask what's the difference between them. 

Is this code incomplete？

1

There is no code related with training, is this code incomplete or there are no plans to open source the training code

deepspeed报错

15

 佬我把sh脚本里的模型改成了没有Instruct过的llama，但是运行torchrun的时候报错了，这个要怎么解决呀

缩写单词合成不准确，有解决方案吗？

1

合成英文的时候在碰到英文单词缩写合成不准确，比如地名缩写NY、BS、KFC等等，请问有解决方案吗？