Commit 24267d0a authored by jameskrw's avatar jameskrw
Browse files

Merge branch 'main' of github.com:JamesKrW/vagen

parents d4de35c4 59af7a91

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

+2 −2
Original line number Diff line number Diff line
@@ -26,7 +26,7 @@ python3 -m vagen.trainer.main_ppo \ 
    data.val_files=data/$EXPERIMENT_NAME/test.parquet \

    data.train_batch_size=64 \

    data.max_prompt_length=1024 \

    data.max_response_length=256 \

    data.max_response_length=128 \

    data.max_trajectory_length=2400 \

    data.image_key=images \

    data.truncation=error \
@@ -67,7 +67,7 @@ python3 -m vagen.trainer.main_ppo \ 
    trainer.experiment_name=$EXPERIMENT_NAME \

    trainer.n_gpus_per_node=4 \

    trainer.nnodes=1 \

    trainer.save_freq=70 \

    trainer.save_freq=90 \

    trainer.test_freq=20 \

    trainer.total_training_steps=200 \

    rollout_manager.max_turns=3 \

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

+2 −2
Original line number Diff line number Diff line
@@ -26,7 +26,7 @@ python3 -m vagen.trainer.main_ppo \ 
    data.val_files=data/$EXPERIMENT_NAME/test.parquet \

    data.train_batch_size=64 \

    data.max_prompt_length=1024 \

    data.max_response_length=256 \

    data.max_response_length=128 \

    data.max_trajectory_length=2400 \

    data.image_key=images \

    data.truncation=error \
@@ -67,7 +67,7 @@ python3 -m vagen.trainer.main_ppo \ 
    trainer.experiment_name=$EXPERIMENT_NAME \

    trainer.n_gpus_per_node=4 \

    trainer.nnodes=1 \

    trainer.save_freq=70 \

    trainer.save_freq=90 \

    trainer.test_freq=20 \

    trainer.total_training_steps=200 \

    rollout_manager.max_turns=3 \

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

+2 −2
Original line number Diff line number Diff line
@@ -26,7 +26,7 @@ python3 -m vagen.trainer.main_ppo \ 
    data.val_files=data/$EXPERIMENT_NAME/test.parquet \

    data.train_batch_size=64 \

    data.max_prompt_length=1024 \

    data.max_response_length=256 \

    data.max_response_length=150 \

    data.max_trajectory_length=2400 \

    data.image_key=images \

    data.truncation=error \
@@ -67,7 +67,7 @@ python3 -m vagen.trainer.main_ppo \ 
    trainer.experiment_name=$EXPERIMENT_NAME \

    trainer.n_gpus_per_node=4 \

    trainer.nnodes=1 \

    trainer.save_freq=70 \

    trainer.save_freq=90 \

    trainer.test_freq=20 \

    trainer.total_training_steps=200 \

    rollout_manager.max_turns=3 \

scripts/exps/format/frozenlake/no_think/run.sh

+2 −2
Original line number Diff line number Diff line
@@ -26,7 +26,7 @@ python3 -m vagen.trainer.main_ppo \ 
    data.val_files=data/$EXPERIMENT_NAME/test.parquet \

    data.train_batch_size=64 \

    data.max_prompt_length=1024 \

    data.max_response_length=256 \

    data.max_response_length=128 \

    data.max_trajectory_length=2400 \

    data.image_key=images \

    data.truncation=error \
@@ -67,7 +67,7 @@ python3 -m vagen.trainer.main_ppo \ 
    trainer.experiment_name=$EXPERIMENT_NAME \

    trainer.n_gpus_per_node=4 \

    trainer.nnodes=1 \

    trainer.save_freq=70 \

    trainer.save_freq=90 \

    trainer.test_freq=20 \

    trainer.total_training_steps=200 \

    rollout_manager.max_turns=3 \

scripts/exps/format/frozenlake/worldmodeling/run.sh

+2 −2
Original line number Diff line number Diff line
@@ -26,7 +26,7 @@ python3 -m vagen.trainer.main_ppo \ 
    data.val_files=data/$EXPERIMENT_NAME/test.parquet \

    data.train_batch_size=64 \

    data.max_prompt_length=1024 \

    data.max_response_length=256 \

    data.max_response_length=128 \

    data.max_trajectory_length=2400 \

    data.image_key=images \

    data.truncation=error \
@@ -67,7 +67,7 @@ python3 -m vagen.trainer.main_ppo \ 
    trainer.experiment_name=$EXPERIMENT_NAME \

    trainer.n_gpus_per_node=4 \

    trainer.nnodes=1 \

    trainer.save_freq=70 \

    trainer.save_freq=90 \

    trainer.test_freq=20 \

    trainer.total_training_steps=200 \

    rollout_manager.max_turns=3 \

CRA Git | Maintained and supported by SUSTech CRA and CCSE