Possible bug that decreases the performance and (maybe) avoids generating the stop token
1
#6 opened 11 months ago
by
davidpeer
Model does not stop generating new tokens.
6
#4 opened 12 months ago
by
MuntasirHossain
Update README.md
2
#3 opened about 1 year ago
by
hadraoui
Repetition from tuning via https://huggingface.co/blog/mlabonne/orpo-llama-3
4
#2 opened about 1 year ago
by
Satya93
Looking forward to full release!
3
4
#1 opened about 1 year ago
by
bartowski