{
  "schema_version": 1,
  "experiment": "7-13",
  "generated_at_utc": "2026-08-03T11:18:21.493471+00:00",
  "real_execution_required": true,
  "upstream_worktree": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream",
  "upstream_commit": "7c51662df27b586f9e8a1ab35fcf849f2b8852f9",
  "arms": {
    "chunk_1": {
      "action_chunk_length": 1,
      "episode_evidence": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_1/episodes.jsonl",
      "rollout_directory": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/rollouts/exp7_13_move_can_pot_chunk_1",
      "environment": {
        "EXP7_13_EPISODE_JSONL": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_1/episodes.jsonl",
        "EXP7_13_ARM": "chunk_1",
        "EXP7_13_UPSTREAM_COMMIT": "7c51662df27b586f9e8a1ab35fcf849f2b8852f9",
        "HYDRA_FULL_ERROR": "1",
        "TOKENIZERS_PARALLELISM": "true",
        "NCCL_DEBUG": "WARN",
        "ROBOT_PLATFORM": "ALOHA",
        "VERL_DISABLE_VLLM_IMPORT": "1"
      },
      "argv": [
        "/home/ubuntu/.venvs/exp7-13-robotwin/bin/python",
        "-u",
        "-m",
        "verl.trainer.main_ppo",
        "data.task_suite_name=robotwin2_move_can_pot",
        "data.num_trials_per_task=128",
        "data.n_samples=1",
        "data.filter_accuracy=False",
        "data.train_batch_size=64",
        "data.val_batch_size=8",
        "data.max_prompt_length=256",
        "data.max_response_length=128",
        "actor_rollout_ref.model.path=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot",
        "actor_rollout_ref.model.vla=openvla-oft",
        "actor_rollout_ref.model.action_token_len=14",
        "actor_rollout_ref.model.action_chunks_len=1",
        "actor_rollout_ref.model.resume=False",
        "actor_rollout_ref.actor.num_images_in_input=3",
        "actor_rollout_ref.actor.traj_mini_batch_size=8",
        "actor_rollout_ref.actor.ppo_mini_batch_size=128",
        "actor_rollout_ref.actor.ppo_micro_batch_size=8",
        "actor_rollout_ref.actor.use_dynamic_bsz=False",
        "actor_rollout_ref.actor.fsdp_config.param_offload=False",
        "actor_rollout_ref.actor.fsdp_config.grad_offload=True",
        "actor_rollout_ref.actor.fsdp_config.optimizer_offload=True",
        "actor_rollout_ref.rollout.twin2_task_config=exp7_13_three_view",
        "actor_rollout_ref.rollout.twin2_instruction_type=seen",
        "actor_rollout_ref.rollout.num_images_in_input=3",
        "+actor_rollout_ref.rollout.action_token_len=14",
        "actor_rollout_ref.rollout.use_proprio=True",
        "actor_rollout_ref.rollout.val_micro_batch_size=8",
        "actor_rollout_ref.rollout.temperature=1.6",
        "actor_rollout_ref.rollout.experiment_name=exp7_13_move_can_pot_chunk_1",
        "actor_rollout_ref.rollout.micro_batch_size=1",
        "actor_rollout_ref.rollout.unnorm_key=robotwin2_move_can_pot_1k",
        "actor_rollout_ref.rollout.model_family=openvla",
        "actor_rollout_ref.rollout.task_suite_name=robotwin2_move_can_pot",
        "actor_rollout_ref.rollout.num_steps_wait=10",
        "actor_rollout_ref.rollout.pretrained_checkpoint=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot",
        "actor_rollout_ref.rollout.center_crop=True",
        "actor_rollout_ref.rollout.max_prompt_length=512",
        "actor_rollout_ref.rollout.tensor_model_parallel_size=1",
        "actor_rollout_ref.rollout.name=hf",
        "actor_rollout_ref.rollout.gpu_memory_utilization=0.9",
        "actor_rollout_ref.ref.log_prob_micro_batch_size=32",
        "actor_rollout_ref.ref.fsdp_config.param_offload=True",
        "algorithm.adv_estimator=grpo",
        "algorithm.kl_ctrl.kl_coef=0.0",
        "trainer.logger=['console']",
        "trainer.project_name=AI-Agent-Book-Experiment-7-13",
        "trainer.experiment_name=exp7_13_move_can_pot_chunk_1",
        "trainer.default_local_dir=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_1/checkpoints",
        "trainer.n_gpus_per_node=1",
        "trainer.nnodes=1",
        "trainer.save_freq=-1",
        "trainer.test_freq=-1",
        "trainer.total_epochs=1",
        "trainer.val_only=True",
        "trainer.val_before_train=True",
        "trainer.runtime_env=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/align.json",
        "trainer.wandb_mode=disabled"
      ],
      "shell_command": "EXP7_13_EPISODE_JSONL=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_1/episodes.jsonl EXP7_13_ARM=chunk_1 EXP7_13_UPSTREAM_COMMIT=7c51662df27b586f9e8a1ab35fcf849f2b8852f9 HYDRA_FULL_ERROR=1 TOKENIZERS_PARALLELISM=true NCCL_DEBUG=WARN ROBOT_PLATFORM=ALOHA VERL_DISABLE_VLLM_IMPORT=1 /home/ubuntu/.venvs/exp7-13-robotwin/bin/python -u -m verl.trainer.main_ppo data.task_suite_name=robotwin2_move_can_pot data.num_trials_per_task=128 data.n_samples=1 data.filter_accuracy=False data.train_batch_size=64 data.val_batch_size=8 data.max_prompt_length=256 data.max_response_length=128 actor_rollout_ref.model.path=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot actor_rollout_ref.model.vla=openvla-oft actor_rollout_ref.model.action_token_len=14 actor_rollout_ref.model.action_chunks_len=1 actor_rollout_ref.model.resume=False actor_rollout_ref.actor.num_images_in_input=3 actor_rollout_ref.actor.traj_mini_batch_size=8 actor_rollout_ref.actor.ppo_mini_batch_size=128 actor_rollout_ref.actor.ppo_micro_batch_size=8 actor_rollout_ref.actor.use_dynamic_bsz=False actor_rollout_ref.actor.fsdp_config.param_offload=False actor_rollout_ref.actor.fsdp_config.grad_offload=True actor_rollout_ref.actor.fsdp_config.optimizer_offload=True actor_rollout_ref.rollout.twin2_task_config=exp7_13_three_view actor_rollout_ref.rollout.twin2_instruction_type=seen actor_rollout_ref.rollout.num_images_in_input=3 actor_rollout_ref.rollout.use_proprio=True actor_rollout_ref.rollout.val_micro_batch_size=8 actor_rollout_ref.rollout.temperature=1.6 actor_rollout_ref.rollout.experiment_name=exp7_13_move_can_pot_chunk_1 actor_rollout_ref.rollout.micro_batch_size=1 actor_rollout_ref.rollout.unnorm_key=robotwin2_move_can_pot_1k actor_rollout_ref.rollout.model_family=openvla actor_rollout_ref.rollout.task_suite_name=robotwin2_move_can_pot actor_rollout_ref.rollout.num_steps_wait=10 actor_rollout_ref.rollout.pretrained_checkpoint=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot actor_rollout_ref.rollout.center_crop=True actor_rollout_ref.rollout.max_prompt_length=512 actor_rollout_ref.rollout.tensor_model_parallel_size=1 actor_rollout_ref.rollout.name=hf actor_rollout_ref.rollout.gpu_memory_utilization=0.9 actor_rollout_ref.ref.log_prob_micro_batch_size=32 actor_rollout_ref.ref.fsdp_config.param_offload=True algorithm.adv_estimator=grpo algorithm.kl_ctrl.kl_coef=0.0 'trainer.logger=['\"'\"'console'\"'\"']' trainer.project_name=AI-Agent-Book-Experiment-7-13 trainer.experiment_name=exp7_13_move_can_pot_chunk_1 trainer.default_local_dir=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_1/checkpoints trainer.n_gpus_per_node=1 trainer.nnodes=1 trainer.save_freq=-1 trainer.test_freq=-1 trainer.total_epochs=1 trainer.val_only=True trainer.val_before_train=True trainer.runtime_env=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/align.json trainer.wandb_mode=disabled"
    },
    "chunk_25": {
      "action_chunk_length": 25,
      "episode_evidence": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_25/episodes.jsonl",
      "rollout_directory": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/rollouts/exp7_13_move_can_pot_chunk_25",
      "environment": {
        "EXP7_13_EPISODE_JSONL": "/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_25/episodes.jsonl",
        "EXP7_13_ARM": "chunk_25",
        "EXP7_13_UPSTREAM_COMMIT": "7c51662df27b586f9e8a1ab35fcf849f2b8852f9",
        "HYDRA_FULL_ERROR": "1",
        "TOKENIZERS_PARALLELISM": "true",
        "NCCL_DEBUG": "WARN",
        "ROBOT_PLATFORM": "ALOHA",
        "VERL_DISABLE_VLLM_IMPORT": "1"
      },
      "argv": [
        "/home/ubuntu/.venvs/exp7-13-robotwin/bin/python",
        "-u",
        "-m",
        "verl.trainer.main_ppo",
        "data.task_suite_name=robotwin2_move_can_pot",
        "data.num_trials_per_task=128",
        "data.n_samples=1",
        "data.filter_accuracy=False",
        "data.train_batch_size=64",
        "data.val_batch_size=8",
        "data.max_prompt_length=256",
        "data.max_response_length=128",
        "actor_rollout_ref.model.path=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot",
        "actor_rollout_ref.model.vla=openvla-oft",
        "actor_rollout_ref.model.action_token_len=14",
        "actor_rollout_ref.model.action_chunks_len=25",
        "actor_rollout_ref.model.resume=False",
        "actor_rollout_ref.actor.num_images_in_input=3",
        "actor_rollout_ref.actor.traj_mini_batch_size=8",
        "actor_rollout_ref.actor.ppo_mini_batch_size=128",
        "actor_rollout_ref.actor.ppo_micro_batch_size=8",
        "actor_rollout_ref.actor.use_dynamic_bsz=False",
        "actor_rollout_ref.actor.fsdp_config.param_offload=False",
        "actor_rollout_ref.actor.fsdp_config.grad_offload=True",
        "actor_rollout_ref.actor.fsdp_config.optimizer_offload=True",
        "actor_rollout_ref.rollout.twin2_task_config=exp7_13_three_view",
        "actor_rollout_ref.rollout.twin2_instruction_type=seen",
        "actor_rollout_ref.rollout.num_images_in_input=3",
        "+actor_rollout_ref.rollout.action_token_len=14",
        "actor_rollout_ref.rollout.use_proprio=True",
        "actor_rollout_ref.rollout.val_micro_batch_size=8",
        "actor_rollout_ref.rollout.temperature=1.6",
        "actor_rollout_ref.rollout.experiment_name=exp7_13_move_can_pot_chunk_25",
        "actor_rollout_ref.rollout.micro_batch_size=1",
        "actor_rollout_ref.rollout.unnorm_key=robotwin2_move_can_pot_1k",
        "actor_rollout_ref.rollout.model_family=openvla",
        "actor_rollout_ref.rollout.task_suite_name=robotwin2_move_can_pot",
        "actor_rollout_ref.rollout.num_steps_wait=10",
        "actor_rollout_ref.rollout.pretrained_checkpoint=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot",
        "actor_rollout_ref.rollout.center_crop=True",
        "actor_rollout_ref.rollout.max_prompt_length=512",
        "actor_rollout_ref.rollout.tensor_model_parallel_size=1",
        "actor_rollout_ref.rollout.name=hf",
        "actor_rollout_ref.rollout.gpu_memory_utilization=0.9",
        "actor_rollout_ref.ref.log_prob_micro_batch_size=32",
        "actor_rollout_ref.ref.fsdp_config.param_offload=True",
        "algorithm.adv_estimator=grpo",
        "algorithm.kl_ctrl.kl_coef=0.0",
        "trainer.logger=['console']",
        "trainer.project_name=AI-Agent-Book-Experiment-7-13",
        "trainer.experiment_name=exp7_13_move_can_pot_chunk_25",
        "trainer.default_local_dir=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_25/checkpoints",
        "trainer.n_gpus_per_node=1",
        "trainer.nnodes=1",
        "trainer.save_freq=-1",
        "trainer.test_freq=-1",
        "trainer.total_epochs=1",
        "trainer.val_only=True",
        "trainer.val_before_train=True",
        "trainer.runtime_env=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/align.json",
        "trainer.wandb_mode=disabled"
      ],
      "shell_command": "EXP7_13_EPISODE_JSONL=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_25/episodes.jsonl EXP7_13_ARM=chunk_25 EXP7_13_UPSTREAM_COMMIT=7c51662df27b586f9e8a1ab35fcf849f2b8852f9 HYDRA_FULL_ERROR=1 TOKENIZERS_PARALLELISM=true NCCL_DEBUG=WARN ROBOT_PLATFORM=ALOHA VERL_DISABLE_VLLM_IMPORT=1 /home/ubuntu/.venvs/exp7-13-robotwin/bin/python -u -m verl.trainer.main_ppo data.task_suite_name=robotwin2_move_can_pot data.num_trials_per_task=128 data.n_samples=1 data.filter_accuracy=False data.train_batch_size=64 data.val_batch_size=8 data.max_prompt_length=256 data.max_response_length=128 actor_rollout_ref.model.path=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot actor_rollout_ref.model.vla=openvla-oft actor_rollout_ref.model.action_token_len=14 actor_rollout_ref.model.action_chunks_len=25 actor_rollout_ref.model.resume=False actor_rollout_ref.actor.num_images_in_input=3 actor_rollout_ref.actor.traj_mini_batch_size=8 actor_rollout_ref.actor.ppo_mini_batch_size=128 actor_rollout_ref.actor.ppo_micro_batch_size=8 actor_rollout_ref.actor.use_dynamic_bsz=False actor_rollout_ref.actor.fsdp_config.param_offload=False actor_rollout_ref.actor.fsdp_config.grad_offload=True actor_rollout_ref.actor.fsdp_config.optimizer_offload=True actor_rollout_ref.rollout.twin2_task_config=exp7_13_three_view actor_rollout_ref.rollout.twin2_instruction_type=seen actor_rollout_ref.rollout.num_images_in_input=3 actor_rollout_ref.rollout.use_proprio=True actor_rollout_ref.rollout.val_micro_batch_size=8 actor_rollout_ref.rollout.temperature=1.6 actor_rollout_ref.rollout.experiment_name=exp7_13_move_can_pot_chunk_25 actor_rollout_ref.rollout.micro_batch_size=1 actor_rollout_ref.rollout.unnorm_key=robotwin2_move_can_pot_1k actor_rollout_ref.rollout.model_family=openvla actor_rollout_ref.rollout.task_suite_name=robotwin2_move_can_pot actor_rollout_ref.rollout.num_steps_wait=10 actor_rollout_ref.rollout.pretrained_checkpoint=/home/ubuntu/checkpoints/exp7-13-openvla-move-can-pot actor_rollout_ref.rollout.center_crop=True actor_rollout_ref.rollout.max_prompt_length=512 actor_rollout_ref.rollout.tensor_model_parallel_size=1 actor_rollout_ref.rollout.name=hf actor_rollout_ref.rollout.gpu_memory_utilization=0.9 actor_rollout_ref.ref.log_prob_micro_batch_size=32 actor_rollout_ref.ref.fsdp_config.param_offload=True algorithm.adv_estimator=grpo algorithm.kl_ctrl.kl_coef=0.0 'trainer.logger=['\"'\"'console'\"'\"']' trainer.project_name=AI-Agent-Book-Experiment-7-13 trainer.experiment_name=exp7_13_move_can_pot_chunk_25 trainer.default_local_dir=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/chunk_25/checkpoints trainer.n_gpus_per_node=1 trainer.nnodes=1 trainer.save_freq=-1 trainer.test_freq=-1 trainer.total_epochs=1 trainer.val_only=True trainer.val_before_train=True trainer.runtime_env=/home/ubuntu/ai-agent-book/chapter7/openvla-robotwin2-eval/runs/exp7-13-localgpu-20260803-v1/instrumented-upstream/align.json trainer.wandb_mode=disabled"
    }
  }
}