mit-han-lab offsite-tuning issues - Githubissues

mit-han-lab / offsite-tuning

Offsite-Tuning: Transfer Learning without Full Model

https://arxiv.org/abs/2302.04870

MIT License

366 stars 37 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

LM loss is nan

#11 QDPeng closed 3 months ago
1
Usage of Pile dataset to train the emulator

#10 ziqi-zhang opened 8 months ago
1
Need to add a Loop here?

#9 talkingwallace opened 1 year ago
0
Evaluation metrics is only perplexity?

#8 Smu-Tan closed 1 year ago
1
Suggestion to upload large files to cloud drive

#7 iseesaw closed 9 months ago
3
NotImplementedError

#6 vis-face opened 1 year ago
1
RuntimeError: expected scalar type Half but found Float

#5 Ryan0v0 closed 1 year ago
1
How to run distributed evaluation for big models used in this paper?

#4 sanyalsunny111 opened 1 year ago
0
The choose_layers_by_changes function is not implemented in the utils.py file.

#3 gray311 closed 1 year ago
3
The authors should consider changing the term “adapter” to avoid potential confusion with adapter-tuning.

#2 henryzhongsc opened 1 year ago
0
fix typo in eva_vit.py

#1 eltociear opened 1 year ago
0