Loading vagen/trainer/config/ppo_trainer.yaml +1 −0 Original line number Diff line number Diff line Loading @@ -31,6 +31,7 @@ actor_rollout_ref: use_dynamic_bsz: False ppo_max_token_len_per_gpu: 16384 # n * ${data.max_prompt_length} + ${data.max_response_length} grad_clip: 1.0 grad_norm_threshold: 10000 clip_ratio: 0.2 entropy_coeff: 0.001 use_kl_loss: False # True for GRPO Loading Loading
vagen/trainer/config/ppo_trainer.yaml +1 −0 Original line number Diff line number Diff line Loading @@ -31,6 +31,7 @@ actor_rollout_ref: use_dynamic_bsz: False ppo_max_token_len_per_gpu: 16384 # n * ${data.max_prompt_length} + ${data.max_response_length} grad_clip: 1.0 grad_norm_threshold: 10000 clip_ratio: 0.2 entropy_coeff: 0.001 use_kl_loss: False # True for GRPO Loading
