Unverified Commit 6405613e authored by Kangrui Wang's avatar Kangrui Wang Committed by GitHub
Browse files

Merge pull request #27 from RAGEN-AI/format 

Format
parents 4a7a1ae4 ee1838ae

scripts/exps/format/maniskill/free_think/run.sh

+2 −2
Original line number Diff line number Diff line
@@ -33,8 +33,8 @@ python3 -m vagen.trainer.main_ppo \ 
    data.train_batch_size=64 \

    data.val_batch_size=32 \

    data.max_prompt_length=1024 \

    data.max_response_length=150 \

    data.max_trajectory_length=3600 \

    data.max_response_length=200 \

    data.max_trajectory_length=3800 \

    data.image_key=images \

    data.truncation=error \

    actor_rollout_ref.model.path=Qwen/Qwen2.5-VL-3B-Instruct \

scripts/exps/format/maniskill/grounding/run.sh

+2 −2
Original line number Diff line number Diff line
@@ -33,8 +33,8 @@ python3 -m vagen.trainer.main_ppo \ 
    data.train_batch_size=64 \

    data.val_batch_size=32 \

    data.max_prompt_length=1024 \

    data.max_response_length=150 \

    data.max_trajectory_length=3600 \

    data.max_response_length=200 \

    data.max_trajectory_length=3800 \

    data.image_key=images \

    data.truncation=error \

    actor_rollout_ref.model.path=Qwen/Qwen2.5-VL-3B-Instruct \

scripts/exps/format/maniskill/grounding/run_tmux.sh

+2 −2
Original line number Diff line number Diff line
@@ -77,8 +77,8 @@ tmux send-keys -t "$TRAIN_SESSION" "python3 -m vagen.trainer.main_ppo \\ 
    data.train_batch_size=64 \\

    data.val_batch_size=32 \\

    data.max_prompt_length=1024 \\

    data.max_response_length=150 \\

    data.max_trajectory_length=3600 \\

    data.max_response_length=200 \\

    data.max_trajectory_length=3800 \\

    data.image_key=images \\

    data.truncation=error \\

    actor_rollout_ref.model.path=Qwen/Qwen2.5-VL-3B-Instruct \\

scripts/exps/format/maniskill/grounding_worldmodeling/run.sh

+2 −2
Original line number Diff line number Diff line
@@ -33,8 +33,8 @@ python3 -m vagen.trainer.main_ppo \ 
    data.train_batch_size=64 \

    data.val_batch_size=32 \

    data.max_prompt_length=1024 \

    data.max_response_length=150 \

    data.max_trajectory_length=3600 \

    data.max_response_length=200 \

    data.max_trajectory_length=3800 \

    data.image_key=images \

    data.truncation=error \

    actor_rollout_ref.model.path=Qwen/Qwen2.5-VL-3B-Instruct \

scripts/exps/format/maniskill/grounding_worldmodeling/run_tmux.sh

+2 −2
Original line number Diff line number Diff line
@@ -77,8 +77,8 @@ tmux send-keys -t "$TRAIN_SESSION" "python3 -m vagen.trainer.main_ppo \\ 
    data.train_batch_size=64 \\

    data.val_batch_size=32 \\

    data.max_prompt_length=1024 \\

    data.max_response_length=150 \\

    data.max_trajectory_length=3600 \\

    data.max_response_length=200 \\

    data.max_trajectory_length=3800 \\

    data.image_key=images \\

    data.truncation=error \\

    actor_rollout_ref.model.path=Qwen/Qwen2.5-VL-3B-Instruct \\

CRA Git | Maintained and supported by SUSTech CRA and CCSE