minor (f1383513) · Commits · 张泽凯 / VAGEN

vagen/examples/release_experiments/mask_gae_mask_loss_turnwise_reward_bi_level.sh

+4 −4

Original line number	Diff line number	Diff line
		@@ -46,7 +46,7 @@ python3 -m vagen.trainer.main_ppo \
		actor_rollout_ref.rollout.log_prob_micro_batch_size_per_gpu=1 \
		actor_rollout_ref.rollout.tensor_model_parallel_size=2 \
		actor_rollout_ref.rollout.name=vllm \
		actor_rollout_ref.rollout.gpu_memory_utilization=0.2 \
		actor_rollout_ref.rollout.gpu_memory_utilization=0.4 \
		actor_rollout_ref.rollout.enable_chunked_prefill=False \
		actor_rollout_ref.rollout.enforce_eager=False \
		actor_rollout_ref.rollout.free_cache_engine=False \
		@@ -67,10 +67,10 @@ python3 -m vagen.trainer.main_ppo \
		trainer.logger=['console','wandb'] \
		trainer.project_name='vagen' \
		trainer.experiment_name='mask_gae_mask_loss_turnwise_reward_bi_level' \
		trainer.n_gpus_per_node=4 \
		trainer.n_gpus_per_node=8 \
		trainer.nnodes=1 \
		trainer.save_freq=100 \
		trainer.test_freq=5 \
		trainer.save_freq=70 \
		trainer.test_freq=20 \
		trainer.total_epochs=15 \
		rollout_manager.max_turns=3 \
		rollout_manager.window_size=5 \