Merge pull request #1553 from hannlp/hans

Change the default argument settings for PPO training Former-commit-id: 1b64678fa4979485f67c3bb1420dfdff6fcbc6e7
2025-12-17 20:30:36 +08:00 · 2023-11-20 20:32:55 +08:00
parent c815b51fef 3f7cdf71ad
commit 05fd97c637
4 changed files with 9 additions and 1 deletions
--- a/README.md
+++ b/README.md
@@ -313,6 +313,8 @@ CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
    --per_device_train_batch_size 2 \
    --gradient_accumulation_steps 4 \
    --lr_scheduler_type cosine \
+    --top_k 0 \
+    --top_p 0.9 \
    --logging_steps 10 \
    --save_steps 1000 \
    --learning_rate 1e-5 \