423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-17 04:10:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix dpo metrics

Browse Source 
Former-commit-id: 57029280da825a39fbf5a05097921b861f126669
This commit is contained in:
hiyouga
2024-11-02 19:22:11 +08:00
parent 174c8fdffd
commit 0ce5b84ccd
7 changed files with 143 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llamafactory/train/kto/workflow.py
View File

@@ -81,7 +81,7 @@ def run_kto(
trainer.save_metrics("train", train_result.metrics)
trainer.save_state()
if trainer.is_world_process_zero() and finetuning_args.plot_loss:
plot_loss(training_args.output_dir, keys=["loss", "eval_loss", "train/rewards/chosen"])
plot_loss(training_args.output_dir, keys=["loss", "eval_loss", "rewards/chosen"])
# Evaluation
if training_args.do_eval: