imyzx2017

【UNIMO】Will you release the trainning code of UNIMO?

1

Q: Will you release the trainning code of UNIMO? Thanks for your reply.

The pretrained STME module

3

Did you re-trained the STME module based on 'attnGAN' ? If so, could you provided it? thanks! After adding my own pretrained caption model, replacing the DAMSM loss to caption_loss...

When I tried gate pruning on gpt2 model (refer to your code). I tried to visulize the gate_value and make one gif, but my result is: all the log_a is...

When I run the 'rbm_chords.py ', an error occurs: module 'midi' has no attribute 'read_midifile', how can I solve it, Thanks!

The losses of G were too large when training

5

  Here is my 3 stages generator's loss changing during training on CUB dataset, all the parameters were used the default setting. During training, the D's loss was decrease,...

I tried implementing R precision by myself, but the result I got on attnGAN model, was very low, near 16.37%.（R=1） So how do you implement the R precision?Could you share...

Thank you for your great work, Here I'm curious about MOE-Transformer's static graph construction. > Q: When there is 1024 experts, switch gating method is used, you need to build...