刘安平
刘安平
When I try to use the provided python script dump_rnn.py to decode the newweights9i.hdf5 model, I found that it can not work well. So I change a lot of it...
### Is there an existing issue for this? - [X] I have searched the existing issues ### Current Behavior 我发现长文本生成效果不容易调出来，特别容易出现重复 例如“我吃饭了吗吗吗吗吗吗吗吗吗”，网上说是退化问题，即随着生成文本长度的增加其质量会逐渐降低，容易出现多种层次（字、短语、句子级）的重复生成。有没有大神给一些有效的经验。 ### Expected Behavior 求大神指点 ### Steps To Reproduce 训练长文本生成。...
i don't see anything new till now. how about big model in audio
13b 训练几十亿token你们用了多少卡 多久
why don't update large model in audio