Upload folder using huggingface_hub

.gitattributes

@@ -427,3 +427,20 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 2026.05.06/04.52.43_train_llmbc_lowdim_box-close-v2/wandb/run-20260506_045327-t6qu9x0g/run-t6qu9x0g.wandb filter=lfs diff=lfs merge=lfs -text
 2026.05.06/05.04.12_train_llmbc_lowdim_box-close-v2/logs.json.txt filter=lfs diff=lfs merge=lfs -text
 2026.05.06/05.04.12_train_llmbc_lowdim_box-close-v2/wandb/run-20260506_050458-k7lt6rm7/run-k7lt6rm7.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/run-ags56sz9.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/wandb/run-20260410_093742-2o6jqedz/run-2o6jqedz.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.42.47_train_llmbc_lowdim_adroit-hand-relocate-v1/logs.json.txt filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.42.47_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_094422-68a5glbf/run-68a5glbf.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.42.50_train_llmbc_lowdim_adroit-hand-relocate-v1/logs.json.txt filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.42.50_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_094424-vk0ow3fl/run-vk0ow3fl.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/09.46.35_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_095519-3qme4js3/run-3qme4js3.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/14.18.53_train_rbc_reward_model_adroit-hand-relocate-v1/wandb/run-20260410_141855-1qgdq8b7/run-1qgdq8b7.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/14.21.55_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_142250-821w1msn/run-821w1msn.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/14.22.41_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_142416-4mgh6loq/run-4mgh6loq.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/14.22.50_train_rbc_reward_model_adroit-hand-relocate-v1/wandb/run-20260410_142252-g9z2m8kf/run-g9z2m8kf.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/18.16.49_train_llmbc_lowdim_adroit-hand-relocate-v1/logs.json.txt filter=lfs diff=lfs merge=lfs -text
+2026.04.10/18.16.49_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_181743-g6hz51in/run-g6hz51in.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/18.20.24_train_llmbc_lowdim_adroit-hand-relocate-v1/logs.json.txt filter=lfs diff=lfs merge=lfs -text
+2026.04.10/18.20.24_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_182424-q1qace9c/run-q1qace9c.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/22.25.42_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_222637-i2nrjl69/run-i2nrjl69.wandb filter=lfs diff=lfs merge=lfs -text
+2026.04.10/22.29.11_train_llmbc_lowdim_adroit-hand-relocate-v1/wandb/run-20260410_223235-ehh65ppa/run-ehh65ppa.wandb filter=lfs diff=lfs merge=lfs -text

2026.04.10/08.02.20_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/08.02.20_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/08.02.20_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/08.02.20_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.02.41_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/09.02.41_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.02.41_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.02.41_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.03.35_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/09.03.35_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.03.35_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.03.35_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.06.09_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/09.06.09_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.06.09_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.06.09_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.06.09_train_bc_lowdim_maze2d-medium-v0/train_bc_lowdim_workspace.log

@@ -0,0 +1,8 @@
+[2026-04-10 09:06:15,647][root][INFO] - running build_ext
+[2026-04-10 09:06:15,649][root][INFO] - building 'mujoco_py.cymj' extension
+[2026-04-10 09:06:15,649][root][INFO] - creating /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py
+[2026-04-10 09:06:15,649][root][INFO] - creating /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl
+[2026-04-10 09:06:15,650][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o -fopenmp -w
+[2026-04-10 09:06:57,299][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -fopenmp -w
+[2026-04-10 09:06:57,388][root][INFO] - creating /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py
+[2026-04-10 09:06:57,389][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -shared -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib -L/venv/llm-bc/lib -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -L/root/.mujoco/mujoco210/bin -Wl,--enable-new-dtags,-rpath,/root/.mujoco/mujoco210/bin -lmujoco210 -lglewosmesa -lOSMesa -lGL -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py/cymj.cpython-39-x86_64-linux-gnu.so -fopenmp

2026.04.10/09.12.25_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/09.12.25_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.12.25_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.12.25_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.12.25_train_bc_lowdim_maze2d-medium-v0/train_bc_lowdim_workspace.log

@@ -0,0 +1,5 @@
+[2026-04-10 09:12:28,251][root][INFO] - running build_ext
+[2026-04-10 09:12:28,253][root][INFO] - building 'mujoco_py.cymj' extension
+[2026-04-10 09:12:28,253][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o -fopenmp -w
+[2026-04-10 09:13:10,017][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -fopenmp -w
+[2026-04-10 09:13:10,101][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -shared -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib -L/venv/llm-bc/lib -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -L/root/.mujoco/mujoco210/bin -Wl,--enable-new-dtags,-rpath,/root/.mujoco/mujoco210/bin -lmujoco210 -lglewosmesa -lOSMesa -lGL -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py/cymj.cpython-39-x86_64-linux-gnu.so -fopenmp

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,121 @@
+name: train_bc_lowdim
+_target_: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+policy:
+  _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 128
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.01
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0.pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train_bc_lowdim_workspace
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: train_bc_lowdim_workspace
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/llmbc/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0
+    choices:
+      task: maze2d-medium-v0
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/logs.json.txt

@@ -0,0 +1,198 @@
+{"train_loss": 0.5742757320404053, "train_loss_bc": 0.5742757320404053, "grad_norm": 0.4610179662704468, "global_step": 0, "epoch": 0, "lr": 0.001}
+{"train_loss": 0.5470483899116516, "train_loss_bc": 0.5470483899116516, "grad_norm": 0.3259114921092987, "global_step": 1, "epoch": 0, "lr": 0.002}
+{"train_loss": 0.5437748432159424, "train_loss_bc": 0.5437748432159424, "grad_norm": 0.3096829354763031, "global_step": 2, "epoch": 0, "lr": 0.003}
+{"train_loss": 0.5680013298988342, "train_loss_bc": 0.5680013298988342, "grad_norm": 0.8777705430984497, "global_step": 3, "epoch": 0, "lr": 0.004}
+{"train_loss": 0.5333088636398315, "train_loss_bc": 0.5333088636398315, "grad_norm": 0.5981258749961853, "global_step": 4, "epoch": 0, "lr": 0.005}
+{"train_loss": 0.5779598951339722, "train_loss_bc": 0.5779598951339722, "grad_norm": 0.5873940587043762, "global_step": 5, "epoch": 0, "lr": 0.006}
+{"train_loss": 0.5605453252792358, "train_loss_bc": 0.5605453252792358, "grad_norm": 0.546937882900238, "global_step": 6, "epoch": 0, "lr": 0.006999999999999999}
+{"train_loss": 0.5720486044883728, "train_loss_bc": 0.5720486044883728, "grad_norm": 0.6765738725662231, "global_step": 7, "epoch": 0, "lr": 0.008}
+{"train_loss": 0.5544816255569458, "train_loss_bc": 0.5544816255569458, "grad_norm": 0.7473601698875427, "global_step": 8, "epoch": 0, "lr": 0.009000000000000001}
+{"train_loss": 0.5382111072540283, "train_loss_bc": 0.5382111072540283, "grad_norm": 0.3417593240737915, "global_step": 9, "epoch": 0, "lr": 0.01}
+{"train_loss": 0.5005457997322083, "train_loss_bc": 0.5005457997322083, "grad_norm": 0.2392510324716568, "global_step": 10, "epoch": 0, "lr": 0.009999999999378116}
+{"train_loss": 0.5228367447853088, "train_loss_bc": 0.5228367447853088, "grad_norm": 0.3644765317440033, "global_step": 11, "epoch": 0, "lr": 0.009999999997512466}
+{"train_loss": 0.5359384417533875, "train_loss_bc": 0.5359384417533875, "grad_norm": 0.2075331062078476, "global_step": 12, "epoch": 0, "lr": 0.009999999994403048}
+{"train_loss": 0.46896055340766907, "train_loss_bc": 0.46896055340766907, "grad_norm": 0.1523733139038086, "global_step": 13, "epoch": 0, "lr": 0.009999999990049864}
+{"train_loss": 0.48494070768356323, "train_loss_bc": 0.48494070768356323, "grad_norm": 0.22689184546470642, "global_step": 14, "epoch": 0, "lr": 0.009999999984452912}
+{"train_loss": 0.5497490167617798, "train_loss_bc": 0.5497490167617798, "grad_norm": 0.4582449495792389, "global_step": 15, "epoch": 0, "lr": 0.009999999977612193}
+{"train_loss": 0.5699339509010315, "train_loss_bc": 0.5699339509010315, "grad_norm": 0.680681049823761, "global_step": 16, "epoch": 0, "lr": 0.009999999969527707}
+{"train_loss": 0.5408633947372437, "train_loss_bc": 0.5408633947372437, "grad_norm": 0.2640772759914398, "global_step": 17, "epoch": 0, "lr": 0.009999999960199455}
+{"train_loss": 0.5184565186500549, "train_loss_bc": 0.5184565186500549, "grad_norm": 0.1633172631263733, "global_step": 18, "epoch": 0, "lr": 0.009999999949627435}
+{"train_loss": 0.5436556935310364, "train_loss_bc": 0.5436556935310364, "grad_norm": 0.1845201849937439, "global_step": 19, "epoch": 0, "lr": 0.009999999937811649}
+{"train_loss": 0.5068360567092896, "train_loss_bc": 0.5068360567092896, "grad_norm": 0.3228622078895569, "global_step": 20, "epoch": 0, "lr": 0.009999999924752095}
+{"train_loss": 0.5451252460479736, "train_loss_bc": 0.5451252460479736, "grad_norm": 0.2506679892539978, "global_step": 21, "epoch": 0, "lr": 0.009999999910448773}
+{"train_loss": 0.5045832395553589, "train_loss_bc": 0.5045832395553589, "grad_norm": 0.44697698950767517, "global_step": 22, "epoch": 0, "lr": 0.009999999894901686}
+{"train_loss": 0.5218442678451538, "train_loss_bc": 0.5218442678451538, "grad_norm": 0.3040887713432312, "global_step": 23, "epoch": 0, "lr": 0.00999999987811083}
+{"train_loss": 0.49454182386398315, "train_loss_bc": 0.49454182386398315, "grad_norm": 0.20201249420642853, "global_step": 24, "epoch": 0, "lr": 0.009999999860076207}
+{"train_loss": 0.5165571570396423, "train_loss_bc": 0.5165571570396423, "grad_norm": 0.15043173730373383, "global_step": 25, "epoch": 0, "lr": 0.00999999984079782}
+{"train_loss": 0.5213630199432373, "train_loss_bc": 0.5213630199432373, "grad_norm": 0.2970457673072815, "global_step": 26, "epoch": 0, "lr": 0.009999999820275662}
+{"train_loss": 0.5392494201660156, "train_loss_bc": 0.5392494201660156, "grad_norm": 0.33821403980255127, "global_step": 27, "epoch": 0, "lr": 0.00999999979850974}
+{"train_loss": 0.507509171962738, "train_loss_bc": 0.507509171962738, "grad_norm": 0.16721631586551666, "global_step": 28, "epoch": 0, "lr": 0.009999999775500049}
+{"train_loss": 0.5158230662345886, "train_loss_bc": 0.5158230662345886, "grad_norm": 0.2735518217086792, "global_step": 29, "epoch": 0, "lr": 0.009999999751246592}
+{"train_loss": 0.5095375180244446, "train_loss_bc": 0.5095375180244446, "grad_norm": 0.23328596353530884, "global_step": 30, "epoch": 0, "lr": 0.009999999725749368}
+{"train_loss": 0.4946969449520111, "train_loss_bc": 0.4946969449520111, "grad_norm": 0.2920290231704712, "global_step": 31, "epoch": 0, "lr": 0.009999999699008378}
+{"train_loss": 0.5286844968795776, "train_loss_bc": 0.5286844968795776, "grad_norm": 0.26879623532295227, "global_step": 32, "epoch": 0, "lr": 0.00999999967102362}
+{"train_loss": 0.5065042972564697, "train_loss_bc": 0.5065042972564697, "grad_norm": 0.3493255376815796, "global_step": 33, "epoch": 0, "lr": 0.009999999641795095}
+{"train_loss": 0.5176247954368591, "train_loss_bc": 0.5176247954368591, "grad_norm": 0.32611459493637085, "global_step": 34, "epoch": 0, "lr": 0.009999999611322802}
+{"train_loss": 0.5595735907554626, "train_loss_bc": 0.5595735907554626, "grad_norm": 0.5473532676696777, "global_step": 35, "epoch": 0, "lr": 0.009999999579606744}
+{"train_loss": 0.5184394121170044, "train_loss_bc": 0.5184394121170044, "grad_norm": 0.33806461095809937, "global_step": 36, "epoch": 0, "lr": 0.009999999546646918}
+{"train_loss": 0.4891864061355591, "train_loss_bc": 0.4891864061355591, "grad_norm": 0.2983148396015167, "global_step": 37, "epoch": 0, "lr": 0.009999999512443325}
+{"train_loss": 0.5090031623840332, "train_loss_bc": 0.5090031623840332, "grad_norm": 0.374706894159317, "global_step": 38, "epoch": 0, "lr": 0.009999999476995964}
+{"train_loss": 0.5417673587799072, "train_loss_bc": 0.5417673587799072, "grad_norm": 0.28812533617019653, "global_step": 39, "epoch": 0, "lr": 0.009999999440304838}
+{"train_loss": 0.5432811975479126, "train_loss_bc": 0.5432811975479126, "grad_norm": 0.4054281711578369, "global_step": 40, "epoch": 0, "lr": 0.009999999402369945}
+{"train_loss": 0.5155792832374573, "train_loss_bc": 0.5155792832374573, "grad_norm": 0.41070905327796936, "global_step": 41, "epoch": 0, "lr": 0.009999999363191285}
+{"train_loss": 0.4796150028705597, "train_loss_bc": 0.4796150028705597, "grad_norm": 0.22966867685317993, "global_step": 42, "epoch": 0, "lr": 0.009999999322768858}
+{"train_loss": 0.5101842880249023, "train_loss_bc": 0.5101842880249023, "grad_norm": 0.3219834268093109, "global_step": 43, "epoch": 0, "lr": 0.009999999281102663}
+{"train_loss": 0.502321720123291, "train_loss_bc": 0.502321720123291, "grad_norm": 0.3314169645309448, "global_step": 44, "epoch": 0, "lr": 0.009999999238192702}
+{"train_loss": 0.4937172830104828, "train_loss_bc": 0.4937172830104828, "grad_norm": 0.17465521395206451, "global_step": 45, "epoch": 0, "lr": 0.009999999194038973}
+{"train_loss": 0.5150852799415588, "train_loss_bc": 0.5150852799415588, "grad_norm": 0.30881091952323914, "global_step": 46, "epoch": 0, "lr": 0.009999999148641478}
+{"train_loss": 0.46151822805404663, "train_loss_bc": 0.46151822805404663, "grad_norm": 0.19222350418567657, "global_step": 47, "epoch": 0, "lr": 0.009999999102000216}
+{"train_loss": 0.5235646963119507, "train_loss_bc": 0.5235646963119507, "grad_norm": 0.17854043841362, "global_step": 48, "epoch": 0, "lr": 0.009999999054115189}
+{"train_loss": 0.47330576181411743, "train_loss_bc": 0.47330576181411743, "grad_norm": 0.186621755361557, "global_step": 49, "epoch": 0, "lr": 0.009999999004986394}
+{"train_loss": 0.4679132103919983, "train_loss_bc": 0.4679132103919983, "grad_norm": 0.3300033211708069, "global_step": 50, "epoch": 0, "lr": 0.009999998954613832}
+{"train_loss": 0.4872625470161438, "train_loss_bc": 0.4872625470161438, "grad_norm": 0.20406848192214966, "global_step": 51, "epoch": 0, "lr": 0.009999998902997503}
+{"train_loss": 0.4838597774505615, "train_loss_bc": 0.4838597774505615, "grad_norm": 0.20800356566905975, "global_step": 52, "epoch": 0, "lr": 0.009999998850137406}
+{"train_loss": 0.48533451557159424, "train_loss_bc": 0.48533451557159424, "grad_norm": 0.2890176475048065, "global_step": 53, "epoch": 0, "lr": 0.009999998796033544}
+{"train_loss": 0.49930649995803833, "train_loss_bc": 0.49930649995803833, "grad_norm": 0.2625446617603302, "global_step": 54, "epoch": 0, "lr": 0.009999998740685917}
+{"train_loss": 0.474785178899765, "train_loss_bc": 0.474785178899765, "grad_norm": 0.2095745950937271, "global_step": 55, "epoch": 0, "lr": 0.00999999868409452}
+{"train_loss": 0.46426108479499817, "train_loss_bc": 0.46426108479499817, "grad_norm": 0.23595686256885529, "global_step": 56, "epoch": 0, "lr": 0.009999998626259356}
+{"train_loss": 0.4779570400714874, "train_loss_bc": 0.4779570400714874, "grad_norm": 0.19150982797145844, "global_step": 57, "epoch": 0, "lr": 0.009999998567180427}
+{"train_loss": 0.5089184641838074, "train_loss_bc": 0.5089184641838074, "grad_norm": 0.2511100172996521, "global_step": 58, "epoch": 0, "lr": 0.00999999850685773}
+{"train_loss": 0.4668119549751282, "train_loss_bc": 0.4668119549751282, "grad_norm": 0.1843072921037674, "global_step": 59, "epoch": 0, "lr": 0.00999999844529127}
+{"train_loss": 0.5078861713409424, "train_loss_bc": 0.5078861713409424, "grad_norm": 0.29048284888267517, "global_step": 60, "epoch": 0, "lr": 0.00999999838248104}
+{"train_loss": 0.4471106231212616, "train_loss_bc": 0.4471106231212616, "grad_norm": 0.24524496495723724, "global_step": 61, "epoch": 0, "lr": 0.009999998318427043}
+{"train_loss": 0.4805048108100891, "train_loss_bc": 0.4805048108100891, "grad_norm": 0.21890461444854736, "global_step": 62, "epoch": 0, "lr": 0.009999998253129281}
+{"train_loss": 0.47960126399993896, "train_loss_bc": 0.47960126399993896, "grad_norm": 0.2846362888813019, "global_step": 63, "epoch": 0, "lr": 0.009999998186587752}
+{"train_loss": 0.4453185200691223, "train_loss_bc": 0.4453185200691223, "grad_norm": 0.2093414068222046, "global_step": 64, "epoch": 0, "lr": 0.009999998118802455}
+{"train_loss": 0.47915151715278625, "train_loss_bc": 0.47915151715278625, "grad_norm": 0.25503861904144287, "global_step": 65, "epoch": 0, "lr": 0.009999998049773393}
+{"train_loss": 0.4890253245830536, "train_loss_bc": 0.4890253245830536, "grad_norm": 0.42124173045158386, "global_step": 66, "epoch": 0, "lr": 0.009999997979500564}
+{"train_loss": 0.5161541104316711, "train_loss_bc": 0.5161541104316711, "grad_norm": 0.28259626030921936, "global_step": 67, "epoch": 0, "lr": 0.00999999790798397}
+{"train_loss": 0.46412068605422974, "train_loss_bc": 0.46412068605422974, "grad_norm": 0.26055908203125, "global_step": 68, "epoch": 0, "lr": 0.009999997835223607}
+{"train_loss": 0.4562651216983795, "train_loss_bc": 0.4562651216983795, "grad_norm": 0.34777477383613586, "global_step": 69, "epoch": 0, "lr": 0.009999997761219478}
+{"train_loss": 0.530666708946228, "train_loss_bc": 0.530666708946228, "grad_norm": 0.32242047786712646, "global_step": 70, "epoch": 0, "lr": 0.009999997685971582}
+{"train_loss": 0.5017932057380676, "train_loss_bc": 0.5017932057380676, "grad_norm": 0.2552233040332794, "global_step": 71, "epoch": 0, "lr": 0.00999999760947992}
+{"train_loss": 0.45941221714019775, "train_loss_bc": 0.45941221714019775, "grad_norm": 0.20706579089164734, "global_step": 72, "epoch": 0, "lr": 0.009999997531744494}
+{"train_loss": 0.49711522459983826, "train_loss_bc": 0.49711522459983826, "grad_norm": 0.3599870800971985, "global_step": 73, "epoch": 0, "lr": 0.0099999974527653}
+{"train_loss": 0.4859696328639984, "train_loss_bc": 0.4859696328639984, "grad_norm": 0.2239435315132141, "global_step": 74, "epoch": 0, "lr": 0.009999997372542337}
+{"train_loss": 0.4816194176673889, "train_loss_bc": 0.4816194176673889, "grad_norm": 0.34854719042778015, "global_step": 75, "epoch": 0, "lr": 0.009999997291075612}
+{"train_loss": 0.4330894649028778, "train_loss_bc": 0.4330894649028778, "grad_norm": 0.2513454854488373, "global_step": 76, "epoch": 0, "lr": 0.009999997208365117}
+{"train_loss": 0.45905041694641113, "train_loss_bc": 0.45905041694641113, "grad_norm": 0.3416292071342468, "global_step": 77, "epoch": 0, "lr": 0.009999997124410858}
+{"train_loss": 0.46681252121925354, "train_loss_bc": 0.46681252121925354, "grad_norm": 0.2973605990409851, "global_step": 78, "epoch": 0, "lr": 0.009999997039212831}
+{"train_loss": 0.4290766716003418, "train_loss_bc": 0.4290766716003418, "grad_norm": 0.4387994706630707, "global_step": 79, "epoch": 0, "lr": 0.009999996952771038}
+{"train_loss": 0.4637358784675598, "train_loss_bc": 0.4637358784675598, "grad_norm": 0.2826111614704132, "global_step": 80, "epoch": 0, "lr": 0.00999999686508548}
+{"train_loss": 0.44887471199035645, "train_loss_bc": 0.44887471199035645, "grad_norm": 0.2915211021900177, "global_step": 81, "epoch": 0, "lr": 0.009999996776156154}
+{"train_loss": 0.4132304787635803, "train_loss_bc": 0.4132304787635803, "grad_norm": 0.31090226769447327, "global_step": 82, "epoch": 0, "lr": 0.009999996685983063}
+{"train_loss": 0.42577099800109863, "train_loss_bc": 0.42577099800109863, "grad_norm": 0.33208024501800537, "global_step": 83, "epoch": 0, "lr": 0.009999996594566206}
+{"train_loss": 0.4459473788738251, "train_loss_bc": 0.4459473788738251, "grad_norm": 0.4792470932006836, "global_step": 84, "epoch": 0, "lr": 0.009999996501905582}
+{"train_loss": 0.4580143094062805, "train_loss_bc": 0.4580143094062805, "grad_norm": 0.43946200609207153, "global_step": 85, "epoch": 0, "lr": 0.00999999640800119}
+{"train_loss": 0.46910616755485535, "train_loss_bc": 0.46910616755485535, "grad_norm": 0.2916896641254425, "global_step": 86, "epoch": 0, "lr": 0.009999996312853034}
+{"train_loss": 0.4336342513561249, "train_loss_bc": 0.4336342513561249, "grad_norm": 0.20134015381336212, "global_step": 87, "epoch": 0, "lr": 0.009999996216461112}
+{"train_loss": 0.44419264793395996, "train_loss_bc": 0.44419264793395996, "grad_norm": 0.2640151083469391, "global_step": 88, "epoch": 0, "lr": 0.009999996118825424}
+{"train_loss": 0.4430444538593292, "train_loss_bc": 0.4430444538593292, "grad_norm": 0.301547646522522, "global_step": 89, "epoch": 0, "lr": 0.009999996019945969}
+{"train_loss": 0.43620699644088745, "train_loss_bc": 0.43620699644088745, "grad_norm": 0.34488314390182495, "global_step": 90, "epoch": 0, "lr": 0.009999995919822749}
+{"train_loss": 0.4108070135116577, "train_loss_bc": 0.4108070135116577, "grad_norm": 0.3344149887561798, "global_step": 91, "epoch": 0, "lr": 0.009999995818455763}
+{"train_loss": 0.4386339783668518, "train_loss_bc": 0.4386339783668518, "grad_norm": 0.2507685124874115, "global_step": 92, "epoch": 0, "lr": 0.00999999571584501}
+{"train_loss": 0.43485045433044434, "train_loss_bc": 0.43485045433044434, "grad_norm": 0.3099294900894165, "global_step": 93, "epoch": 0, "lr": 0.00999999561199049}
+{"train_loss": 0.462785005569458, "train_loss_bc": 0.462785005569458, "grad_norm": 0.22741125524044037, "global_step": 94, "epoch": 0, "lr": 0.009999995506892206}
+{"train_loss": 0.4141497015953064, "train_loss_bc": 0.4141497015953064, "grad_norm": 0.22992584109306335, "global_step": 95, "epoch": 0, "lr": 0.009999995400550157}
+{"train_loss": 0.40021058917045593, "train_loss_bc": 0.40021058917045593, "grad_norm": 0.23979420959949493, "global_step": 96, "epoch": 0, "lr": 0.00999999529296434}
+{"train_loss": 0.3920753002166748, "train_loss_bc": 0.3920753002166748, "grad_norm": 0.29057979583740234, "global_step": 97, "epoch": 0, "lr": 0.009999995184134757}
+{"train_loss": 0.4283364415168762, "train_loss_bc": 0.4283364415168762, "grad_norm": 0.5014083981513977, "global_step": 98, "epoch": 0, "lr": 0.00999999507406141}
+{"train_loss": 0.4140693247318268, "train_loss_bc": 0.4140693247318268, "grad_norm": 0.39413923025131226, "global_step": 99, "epoch": 0, "lr": 0.009999994962744294}
+{"train_loss": 0.44777703285217285, "train_loss_bc": 0.44777703285217285, "grad_norm": 0.2801749110221863, "global_step": 100, "epoch": 0, "lr": 0.009999994850183417}
+{"train_loss": 0.4148218631744385, "train_loss_bc": 0.4148218631744385, "grad_norm": 0.40548500418663025, "global_step": 101, "epoch": 0, "lr": 0.00999999473637877}
+{"train_loss": 0.39296433329582214, "train_loss_bc": 0.39296433329582214, "grad_norm": 0.29740381240844727, "global_step": 102, "epoch": 0, "lr": 0.009999994621330359}
+{"train_loss": 0.4314180016517639, "train_loss_bc": 0.4314180016517639, "grad_norm": 0.3907894194126129, "global_step": 103, "epoch": 0, "lr": 0.009999994505038181}
+{"train_loss": 0.42784494161605835, "train_loss_bc": 0.42784494161605835, "grad_norm": 0.434723824262619, "global_step": 104, "epoch": 0, "lr": 0.009999994387502239}
+{"train_loss": 0.4203360080718994, "train_loss_bc": 0.4203360080718994, "grad_norm": 0.41060370206832886, "global_step": 105, "epoch": 0, "lr": 0.00999999426872253}
+{"train_loss": 0.45144379138946533, "train_loss_bc": 0.45144379138946533, "grad_norm": 0.27415886521339417, "global_step": 106, "epoch": 0, "lr": 0.009999994148699056}
+{"train_loss": 0.4519392251968384, "train_loss_bc": 0.4519392251968384, "grad_norm": 0.4237353801727295, "global_step": 107, "epoch": 0, "lr": 0.009999994027431817}
+{"train_loss": 0.3734853267669678, "train_loss_bc": 0.3734853267669678, "grad_norm": 0.2555156350135803, "global_step": 108, "epoch": 0, "lr": 0.009999993904920812}
+{"train_loss": 0.4230925738811493, "train_loss_bc": 0.4230925738811493, "grad_norm": 0.4065951704978943, "global_step": 109, "epoch": 0, "lr": 0.009999993781166042}
+{"train_loss": 0.3788265585899353, "train_loss_bc": 0.3788265585899353, "grad_norm": 0.27832919359207153, "global_step": 110, "epoch": 0, "lr": 0.009999993656167506}
+{"train_loss": 0.4092004895210266, "train_loss_bc": 0.4092004895210266, "grad_norm": 0.2456955760717392, "global_step": 111, "epoch": 0, "lr": 0.009999993529925205}
+{"train_loss": 0.40052908658981323, "train_loss_bc": 0.40052908658981323, "grad_norm": 0.3523780405521393, "global_step": 112, "epoch": 0, "lr": 0.009999993402439136}
+{"train_loss": 0.4118815064430237, "train_loss_bc": 0.4118815064430237, "grad_norm": 0.3752630352973938, "global_step": 113, "epoch": 0, "lr": 0.009999993273709304}
+{"train_loss": 0.39718833565711975, "train_loss_bc": 0.39718833565711975, "grad_norm": 0.33827918767929077, "global_step": 114, "epoch": 0, "lr": 0.009999993143735706}
+{"train_loss": 0.4095107316970825, "train_loss_bc": 0.4095107316970825, "grad_norm": 0.3141377568244934, "global_step": 115, "epoch": 0, "lr": 0.009999993012518343}
+{"train_loss": 0.43222084641456604, "train_loss_bc": 0.43222084641456604, "grad_norm": 0.34126782417297363, "global_step": 116, "epoch": 0, "lr": 0.009999992880057214}
+{"train_loss": 0.39195510745048523, "train_loss_bc": 0.39195510745048523, "grad_norm": 0.358022540807724, "global_step": 117, "epoch": 0, "lr": 0.00999999274635232}
+{"train_loss": 0.4415290355682373, "train_loss_bc": 0.4415290355682373, "grad_norm": 0.3781397342681885, "global_step": 118, "epoch": 0, "lr": 0.009999992611403661}
+{"train_loss": 0.40648651123046875, "train_loss_bc": 0.40648651123046875, "grad_norm": 0.35686495900154114, "global_step": 119, "epoch": 0, "lr": 0.009999992475211238}
+{"train_loss": 0.4031302332878113, "train_loss_bc": 0.4031302332878113, "grad_norm": 0.3705500662326813, "global_step": 120, "epoch": 0, "lr": 0.009999992337775048}
+{"train_loss": 0.45374390482902527, "train_loss_bc": 0.45374390482902527, "grad_norm": 0.2800765037536621, "global_step": 121, "epoch": 0, "lr": 0.009999992199095094}
+{"train_loss": 0.41550350189208984, "train_loss_bc": 0.41550350189208984, "grad_norm": 0.3643583357334137, "global_step": 122, "epoch": 0, "lr": 0.009999992059171374}
+{"train_loss": 0.42902517318725586, "train_loss_bc": 0.42902517318725586, "grad_norm": 0.2818589210510254, "global_step": 123, "epoch": 0, "lr": 0.00999999191800389}
+{"train_loss": 0.3701103925704956, "train_loss_bc": 0.3701103925704956, "grad_norm": 0.28646788001060486, "global_step": 124, "epoch": 0, "lr": 0.00999999177559264}
+{"train_loss": 0.4084678888320923, "train_loss_bc": 0.4084678888320923, "grad_norm": 0.3128051459789276, "global_step": 125, "epoch": 0, "lr": 0.009999991631937623}
+{"train_loss": 0.43478596210479736, "train_loss_bc": 0.43478596210479736, "grad_norm": 0.2467166632413864, "global_step": 126, "epoch": 0, "lr": 0.009999991487038845}
+{"train_loss": 0.4194517135620117, "train_loss_bc": 0.4194517135620117, "grad_norm": 0.1867067962884903, "global_step": 127, "epoch": 0, "lr": 0.009999991340896299}
+{"train_loss": 0.40881070494651794, "train_loss_bc": 0.40881070494651794, "grad_norm": 0.4022925794124603, "global_step": 128, "epoch": 0, "lr": 0.00999999119350999}
+{"train_loss": 0.41983669996261597, "train_loss_bc": 0.41983669996261597, "grad_norm": 0.27260592579841614, "global_step": 129, "epoch": 0, "lr": 0.009999991044879915}
+{"train_loss": 0.4503186345100403, "train_loss_bc": 0.4503186345100403, "grad_norm": 0.40697935223579407, "global_step": 130, "epoch": 0, "lr": 0.009999990895006076}
+{"train_loss": 0.4265207052230835, "train_loss_bc": 0.4265207052230835, "grad_norm": 0.2521716356277466, "global_step": 131, "epoch": 0, "lr": 0.009999990743888474}
+{"train_loss": 0.4375496506690979, "train_loss_bc": 0.4375496506690979, "grad_norm": 0.27460354566574097, "global_step": 132, "epoch": 0, "lr": 0.009999990591527104}
+{"train_loss": 0.4457821846008301, "train_loss_bc": 0.4457821846008301, "grad_norm": 0.24064527451992035, "global_step": 133, "epoch": 0, "lr": 0.009999990437921972}
+{"train_loss": 0.4605506360530853, "train_loss_bc": 0.4605506360530853, "grad_norm": 0.2716171145439148, "global_step": 134, "epoch": 0, "lr": 0.009999990283073072}
+{"train_loss": 0.38944196701049805, "train_loss_bc": 0.38944196701049805, "grad_norm": 0.30115771293640137, "global_step": 135, "epoch": 0, "lr": 0.009999990126980409}
+{"train_loss": 0.41354790329933167, "train_loss_bc": 0.41354790329933167, "grad_norm": 0.33541253209114075, "global_step": 136, "epoch": 0, "lr": 0.009999989969643982}
+{"train_loss": 0.37190109491348267, "train_loss_bc": 0.37190109491348267, "grad_norm": 0.19214670360088348, "global_step": 137, "epoch": 0, "lr": 0.00999998981106379}
+{"train_loss": 0.4145277738571167, "train_loss_bc": 0.4145277738571167, "grad_norm": 0.3911716043949127, "global_step": 138, "epoch": 0, "lr": 0.009999989651239834}
+{"train_loss": 0.43325337767601013, "train_loss_bc": 0.43325337767601013, "grad_norm": 0.3034513592720032, "global_step": 139, "epoch": 0, "lr": 0.009999989490172112}
+{"train_loss": 0.36040061712265015, "train_loss_bc": 0.36040061712265015, "grad_norm": 0.2691652178764343, "global_step": 140, "epoch": 0, "lr": 0.009999989327860628}
+{"train_loss": 0.3998282551765442, "train_loss_bc": 0.3998282551765442, "grad_norm": 0.2916651666164398, "global_step": 141, "epoch": 0, "lr": 0.009999989164305377}
+{"train_loss": 0.4111650586128235, "train_loss_bc": 0.4111650586128235, "grad_norm": 0.27103137969970703, "global_step": 142, "epoch": 0, "lr": 0.009999988999506364}
+{"train_loss": 0.36338406801223755, "train_loss_bc": 0.36338406801223755, "grad_norm": 0.2345210313796997, "global_step": 143, "epoch": 0, "lr": 0.009999988833463584}
+{"train_loss": 0.3807051479816437, "train_loss_bc": 0.3807051479816437, "grad_norm": 0.2513706684112549, "global_step": 144, "epoch": 0, "lr": 0.009999988666177042}
+{"train_loss": 0.4108376204967499, "train_loss_bc": 0.4108376204967499, "grad_norm": 0.323003888130188, "global_step": 145, "epoch": 0, "lr": 0.009999988497646734}
+{"train_loss": 0.3920934796333313, "train_loss_bc": 0.3920934796333313, "grad_norm": 0.26234710216522217, "global_step": 146, "epoch": 0, "lr": 0.009999988327872664}
+{"train_loss": 0.3924039304256439, "train_loss_bc": 0.3924039304256439, "grad_norm": 0.3997865617275238, "global_step": 147, "epoch": 0, "lr": 0.009999988156854829}
+{"train_loss": 0.3870144784450531, "train_loss_bc": 0.3870144784450531, "grad_norm": 0.27861636877059937, "global_step": 148, "epoch": 0, "lr": 0.00999998798459323}
+{"train_loss": 0.38959047198295593, "train_loss_bc": 0.38959047198295593, "grad_norm": 0.3393172025680542, "global_step": 149, "epoch": 0, "lr": 0.009999987811087865}
+{"train_loss": 0.40116578340530396, "train_loss_bc": 0.40116578340530396, "grad_norm": 0.2975669503211975, "global_step": 150, "epoch": 0, "lr": 0.009999987636338739}
+{"train_loss": 0.42044979333877563, "train_loss_bc": 0.42044979333877563, "grad_norm": 0.44693636894226074, "global_step": 151, "epoch": 0, "lr": 0.009999987460345847}
+{"train_loss": 0.35583871603012085, "train_loss_bc": 0.35583871603012085, "grad_norm": 0.34289121627807617, "global_step": 152, "epoch": 0, "lr": 0.009999987283109192}
+{"train_loss": 0.41687560081481934, "train_loss_bc": 0.41687560081481934, "grad_norm": 0.44672927260398865, "global_step": 153, "epoch": 0, "lr": 0.009999987104628773}
+{"train_loss": 0.40855151414871216, "train_loss_bc": 0.40855151414871216, "grad_norm": 0.2894774377346039, "global_step": 154, "epoch": 0, "lr": 0.00999998692490459}
+{"train_loss": 0.4261748790740967, "train_loss_bc": 0.4261748790740967, "grad_norm": 0.37578925490379333, "global_step": 155, "epoch": 0, "lr": 0.009999986743936642}
+{"train_loss": 0.41958966851234436, "train_loss_bc": 0.41958966851234436, "grad_norm": 0.39029765129089355, "global_step": 156, "epoch": 0, "lr": 0.009999986561724931}
+{"train_loss": 0.4157213270664215, "train_loss_bc": 0.4157213270664215, "grad_norm": 0.550552487373352, "global_step": 157, "epoch": 0, "lr": 0.009999986378269458}
+{"train_loss": 0.3687168061733246, "train_loss_bc": 0.3687168061733246, "grad_norm": 0.32177433371543884, "global_step": 158, "epoch": 0, "lr": 0.00999998619357022}
+{"train_loss": 0.37296295166015625, "train_loss_bc": 0.37296295166015625, "grad_norm": 0.7262616753578186, "global_step": 159, "epoch": 0, "lr": 0.009999986007627217}
+{"train_loss": 0.40853071212768555, "train_loss_bc": 0.40853071212768555, "grad_norm": 0.7078126668930054, "global_step": 160, "epoch": 0, "lr": 0.009999985820440453}
+{"train_loss": 0.42252951860427856, "train_loss_bc": 0.42252951860427856, "grad_norm": 0.34911733865737915, "global_step": 161, "epoch": 0, "lr": 0.009999985632009925}
+{"train_loss": 0.42843812704086304, "train_loss_bc": 0.42843812704086304, "grad_norm": 0.2839755415916443, "global_step": 162, "epoch": 0, "lr": 0.009999985442335633}
+{"train_loss": 0.37448471784591675, "train_loss_bc": 0.37448471784591675, "grad_norm": 0.3231787383556366, "global_step": 163, "epoch": 0, "lr": 0.009999985251417575}
+{"train_loss": 0.38526156544685364, "train_loss_bc": 0.38526156544685364, "grad_norm": 0.42802849411964417, "global_step": 164, "epoch": 0, "lr": 0.009999985059255758}
+{"train_loss": 0.43893128633499146, "train_loss_bc": 0.43893128633499146, "grad_norm": 0.43370962142944336, "global_step": 165, "epoch": 0, "lr": 0.009999984865850176}
+{"train_loss": 0.36541861295700073, "train_loss_bc": 0.36541861295700073, "grad_norm": 0.3593718707561493, "global_step": 166, "epoch": 0, "lr": 0.009999984671200831}
+{"train_loss": 0.37354975938796997, "train_loss_bc": 0.37354975938796997, "grad_norm": 0.31158536672592163, "global_step": 167, "epoch": 0, "lr": 0.00999998447530772}
+{"train_loss": 0.3358023166656494, "train_loss_bc": 0.3358023166656494, "grad_norm": 0.3619420826435089, "global_step": 168, "epoch": 0, "lr": 0.00999998427817085}
+{"train_loss": 0.3332926630973816, "train_loss_bc": 0.3332926630973816, "grad_norm": 0.2583524286746979, "global_step": 169, "epoch": 0, "lr": 0.009999984079790214}
+{"train_loss": 0.3937195837497711, "train_loss_bc": 0.3937195837497711, "grad_norm": 0.4220624566078186, "global_step": 170, "epoch": 0, "lr": 0.009999983880165816}
+{"train_loss": 0.3483230471611023, "train_loss_bc": 0.3483230471611023, "grad_norm": 0.2745370864868164, "global_step": 171, "epoch": 0, "lr": 0.009999983679297653}
+{"train_loss": 0.3706860840320587, "train_loss_bc": 0.3706860840320587, "grad_norm": 0.3540709912776947, "global_step": 172, "epoch": 0, "lr": 0.00999998347718573}
+{"train_loss": 0.3873794674873352, "train_loss_bc": 0.3873794674873352, "grad_norm": 0.3056386113166809, "global_step": 173, "epoch": 0, "lr": 0.009999983273830042}
+{"train_loss": 0.3573632538318634, "train_loss_bc": 0.3573632538318634, "grad_norm": 0.4181464910507202, "global_step": 174, "epoch": 0, "lr": 0.009999983069230592}
+{"train_loss": 0.42982178926467896, "train_loss_bc": 0.42982178926467896, "grad_norm": 0.4089113473892212, "global_step": 175, "epoch": 0, "lr": 0.00999998286338738}
+{"train_loss": 0.3415180444717407, "train_loss_bc": 0.3415180444717407, "grad_norm": 0.2808316946029663, "global_step": 176, "epoch": 0, "lr": 0.009999982656300404}
+{"train_loss": 0.39931759238243103, "train_loss_bc": 0.39931759238243103, "grad_norm": 0.31670093536376953, "global_step": 177, "epoch": 0, "lr": 0.009999982447969664}
+{"train_loss": 0.42158377170562744, "train_loss_bc": 0.42158377170562744, "grad_norm": 0.434965580701828, "global_step": 178, "epoch": 0, "lr": 0.009999982238395164}
+{"train_loss": 0.35083356499671936, "train_loss_bc": 0.35083356499671936, "grad_norm": 0.4147171676158905, "global_step": 179, "epoch": 0, "lr": 0.0099999820275769}
+{"train_loss": 0.34347087144851685, "train_loss_bc": 0.34347087144851685, "grad_norm": 0.317233145236969, "global_step": 180, "epoch": 0, "lr": 0.009999981815514874}
+{"train_loss": 0.3805704712867737, "train_loss_bc": 0.3805704712867737, "grad_norm": 0.2734706997871399, "global_step": 181, "epoch": 0, "lr": 0.009999981602209084}
+{"train_loss": 0.3702990710735321, "train_loss_bc": 0.3702990710735321, "grad_norm": 0.36316049098968506, "global_step": 182, "epoch": 0, "lr": 0.009999981387659533}
+{"train_loss": 0.35707738995552063, "train_loss_bc": 0.35707738995552063, "grad_norm": 0.30813276767730713, "global_step": 183, "epoch": 0, "lr": 0.009999981171866219}
+{"train_loss": 0.36711764335632324, "train_loss_bc": 0.36711764335632324, "grad_norm": 0.38487568497657776, "global_step": 184, "epoch": 0, "lr": 0.009999980954829145}
+{"train_loss": 0.34280383586883545, "train_loss_bc": 0.34280383586883545, "grad_norm": 0.277854323387146, "global_step": 185, "epoch": 0, "lr": 0.009999980736548305}
+{"train_loss": 0.33399534225463867, "train_loss_bc": 0.33399534225463867, "grad_norm": 0.4292750358581543, "global_step": 186, "epoch": 0, "lr": 0.009999980517023705}
+{"train_loss": 0.3977174758911133, "train_loss_bc": 0.3977174758911133, "grad_norm": 0.30483928322792053, "global_step": 187, "epoch": 0, "lr": 0.009999980296255342}
+{"train_loss": 0.34298792481422424, "train_loss_bc": 0.34298792481422424, "grad_norm": 0.2111782729625702, "global_step": 188, "epoch": 0, "lr": 0.009999980074243216}
+{"train_loss": 0.32223907113075256, "train_loss_bc": 0.32223907113075256, "grad_norm": 0.29651424288749695, "global_step": 189, "epoch": 0, "lr": 0.009999979850987328}
+{"train_loss": 0.3656037747859955, "train_loss_bc": 0.3656037747859955, "grad_norm": 0.3420378267765045, "global_step": 190, "epoch": 0, "lr": 0.00999997962648768}
+{"train_loss": 0.3618483543395996, "train_loss_bc": 0.3618483543395996, "grad_norm": 0.30036675930023193, "global_step": 191, "epoch": 0, "lr": 0.009999979400744268}
+{"train_loss": 0.38904881477355957, "train_loss_bc": 0.38904881477355957, "grad_norm": 0.28105607628822327, "global_step": 192, "epoch": 0, "lr": 0.009999979173757096}
+{"train_loss": 0.3561739921569824, "train_loss_bc": 0.3561739921569824, "grad_norm": 0.23257242143154144, "global_step": 193, "epoch": 0, "lr": 0.00999997894552616}
+{"train_loss": 0.4250050485134125, "train_loss_bc": 0.4250050485134125, "grad_norm": 0.4045962989330292, "global_step": 194, "epoch": 0, "lr": 0.009999978716051463}
+{"train_loss": 0.36044198274612427, "train_loss_bc": 0.36044198274612427, "grad_norm": 0.3454419672489166, "global_step": 195, "epoch": 0, "lr": 0.009999978485333006}
+{"train_loss": 0.36355459690093994, "train_loss_bc": 0.36355459690093994, "grad_norm": 0.28779804706573486, "global_step": 196, "epoch": 0, "lr": 0.009999978253370786}
+{"train_loss": 0.36438101530075073, "train_loss_bc": 0.36438101530075073, "grad_norm": 0.3752126395702362, "global_step": 197, "epoch": 0, "lr": 0.009999978020164803}

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/train_bc_lowdim_workspace.log

@@ -0,0 +1,11 @@
+[2026-04-10 09:16:09,258][root][INFO] - running build_ext
+[2026-04-10 09:16:09,260][root][INFO] - building 'mujoco_py.cymj' extension
+[2026-04-10 09:16:09,261][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o -fopenmp -w
+[2026-04-10 09:16:51,053][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -fopenmp -w
+[2026-04-10 09:16:51,130][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -shared -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib -L/venv/llm-bc/lib -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -L/root/.mujoco/mujoco210/bin -Wl,--enable-new-dtags,-rpath,/root/.mujoco/mujoco210/bin -lmujoco210 -lglewosmesa -lOSMesa -lGL -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py/cymj.cpython-39-x86_64-linux-gnu.so -fopenmp
+[2026-04-10 09:16:51,979][matplotlib.font_manager][INFO] - generated new fontManager
+[2026-04-10 09:16:56,008][absl][INFO] - MUJOCO_GL=osmesa, attempting to import specified OpenGL backend.
+[2026-04-10 09:16:56,014][absl][INFO] - MuJoCo library version is: 2.3.7
+[2026-04-10 09:16:56,122][datasets][INFO] - PyTorch version 2.2.2 available.
+[2026-04-10 09:16:56,123][datasets][INFO] - TensorFlow version 2.15.1 available.
+[2026-04-10 09:16:56,123][datasets][INFO] - JAX version 0.4.30 available.

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/files/config.yaml

@@ -0,0 +1,199 @@
+_target_:
+    value: llmbc.workspace.train_bc_lowdim_workspace.TrainBCLowdimWorkspace
+_wandb:
+    value:
+        cli_version: 0.18.6
+        m: []
+        python_version: 3.9.25
+        t:
+            "1":
+                - 1
+                - 2
+                - 3
+                - 5
+                - 11
+                - 12
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 55
+                - 71
+                - 83
+                - 95
+                - 100
+                - 105
+            "2":
+                - 1
+                - 2
+                - 3
+                - 5
+                - 11
+                - 12
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 55
+                - 71
+                - 83
+                - 95
+                - 100
+                - 105
+            "3":
+                - 13
+                - 15
+                - 16
+                - 23
+                - 55
+                - 61
+            "4": 3.9.25
+            "5": 0.18.6
+            "6": 4.47.1
+            "8":
+                - 5
+            "12": 0.18.6
+            "13": linux-x86_64
+action_dim:
+    value: 2
+checkpoint:
+    value:
+        save_last_ckpt: true
+        save_last_snapshot: false
+        topk:
+            format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+            k: 5
+            mode: max
+            monitor_key: test_success_rate
+dataloader:
+    value:
+        batch_size: 128
+        num_workers: 0
+        persistent_workers: false
+        pin_memory: false
+        shuffle: true
+exp_name:
+    value: default
+horizon:
+    value: 1
+logging:
+    value:
+        group: null
+        id: null
+        mode: online
+        name: 2026.04.10-09.16.56_train_bc_lowdim_maze2d-medium-v0
+        project: maze2d-medium-v0-training
+        resume: true
+        tags:
+            - train_bc_lowdim
+            - maze2d-medium-v0
+            - default
+multi_run:
+    value:
+        run_dir: data/outputs/2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0
+        wandb_name_base: 2026.04.10-09.16.56_train_bc_lowdim_maze2d-medium-v0
+n_action_steps:
+    value: 1
+n_latency_steps:
+    value: 0
+n_obs_steps:
+    value: 1
+name:
+    value: train_bc_lowdim
+obs_dim:
+    value: 6
+optimizer:
+    value:
+        _target_: torch.optim.AdamW
+        betas:
+            - 0.95
+            - 0.999
+        eps: 1e-08
+        lr: 0.01
+        weight_decay: 1e-06
+output_dir:
+    value: /workspace/LLM-BC/data/outputs/2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0
+past_action_visible:
+    value: false
+policy:
+    value:
+        _target_: llmbc.policy.bc_lowdim_policy.BCLowdimPolicy
+        action_dim: 2
+        horizon: 1
+        model:
+            _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+            activation: relu
+            hidden_size:
+                - 256
+                - 256
+            input_size: 6
+            n_action_steps: 1
+            n_obs_steps: 1
+            output_size: 2
+        n_action_steps: 1
+        n_obs_steps: 1
+        obs_dim: 6
+task:
+    value:
+        action_dim: 2
+        dataset:
+            _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+            data_path: datasets/maze2d-medium-v0.pt
+            data_path2: datasets/maze2d-medium-v0.pt
+            dummy_normalizer: false
+            horizon: 1
+            obs_eef_target: true
+            pad_after: 0
+            pad_before: 0
+            use_manual_normalizer: false
+            val_ratio: 0.02
+        env_runner:
+            _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+            discount: 0.99
+            env_name: llf-pointmaze-maze2d-medium-v0
+            feedback_type:
+                - hp
+                - hn
+                - fp
+            instruction_type: b
+            max_steps: 250
+            n_action_steps: 1
+            n_envs: 10
+            n_obs_steps: 1
+            n_test: 50
+            n_train: 10
+            visual: false
+        instructor:
+            _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor
+        name: maze2d-medium-v0
+        obs_dim: 6
+task_name:
+    value: maze2d-medium-v0
+training:
+    value:
+        checkpoint_every: 5
+        debug: false
+        device: cuda:0
+        grad_norm_clip: 0.5
+        gradient_accumulate_every: 1
+        lr_scheduler: cosine
+        lr_warmup_steps: 10
+        max_train_steps: null
+        max_val_steps: null
+        num_epochs: 1001
+        resume: false
+        rollout_every: 5
+        sample_every: 5
+        sample_max_batch: 128
+        seed: 42
+        tqdm_interval_sec: 1
+        val_every: 1
+val_dataloader:
+    value:
+        batch_size: 128
+        num_workers: 0
+        persistent_workers: false
+        pin_memory: false
+        shuffle: true

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/files/output.log

@@ -0,0 +1,60 @@
+/venv/llm-bc/lib/python3.9/site-packages/gym/envs/registration.py:505: UserWarning: WARN: The environment maze2d-medium-v0 is out of date. You should consider upgrading to version `v1` with the environment ID `maze2d-medium-v1`.
+  logger.warn(
+Eval PointmazeLowdimRunner 1/6:   0%|                                                                                                                                   | 0/250 [00:00<?, ?it/s]/workspace/LLM-BC/llmbc/common/llfbench_util.py:39: UserWarning: Creating a tensor from a list of numpy.ndarrays is extremely slow. Please consider converting the list to a single numpy.ndarray with numpy.array() before converting to a tensor. (Triggered internally at ../torch/csrc/utils/tensor_new.cpp:275.)
+  obs = torch.tensor(obs, dtype=torch.float32).unsqueeze(dim=0).to(device)
+Eval PointmazeLowdimRunner 2/6:  54%|████████████████████████████████████████████████████████████████▊                                                        | 134/250 [00:04<00:03, 36.40it/s]Traceback (most recent call last):
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 283, in <module>
+    main()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/main.py", line 90, in decorated_main
+    _run_hydra(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 389, in _run_hydra
+    _run_app(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 452, in _run_app
+    run_and_report(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 213, in run_and_report
+    return func()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 453, in <lambda>
+    lambda: hydra.run(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/hydra.py", line 119, in run
+    ret = run_job(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/core/utils.py", line 186, in run_job
+    ret.return_value = task_function(task_cfg)
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 280, in main
+    workspace.run()
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 199, in run
+    runner_log = env_runner.run(policy)
+  File "/workspace/LLM-BC/llmbc/env_runner/pointmaze_lowdim_runner.py", line 149, in run
+    obs = obs_converter(observation, visual=self.visual)
+  File "/workspace/LLM-BC/llmbc/common/llfbench_util.py", line 10, in obs_converter
+    ob = obs_converter(obs_v, visual, obs_image)
+  File "/workspace/LLM-BC/llmbc/common/llfbench_util.py", line 39, in obs_converter
+    obs = torch.tensor(obs, dtype=torch.float32).unsqueeze(dim=0).to(device)
+KeyboardInterrupt
+Traceback (most recent call last):
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 283, in <module>
+    main()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/main.py", line 90, in decorated_main
+    _run_hydra(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 389, in _run_hydra
+    _run_app(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 452, in _run_app
+    run_and_report(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 213, in run_and_report
+    return func()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 453, in <lambda>
+    lambda: hydra.run(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/hydra.py", line 119, in run
+    ret = run_job(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/core/utils.py", line 186, in run_job
+    ret.return_value = task_function(task_cfg)
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 280, in main
+    workspace.run()
+  File "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py", line 199, in run
+    runner_log = env_runner.run(policy)
+  File "/workspace/LLM-BC/llmbc/env_runner/pointmaze_lowdim_runner.py", line 149, in run
+    obs = obs_converter(observation, visual=self.visual)
+  File "/workspace/LLM-BC/llmbc/common/llfbench_util.py", line 10, in obs_converter
+    ob = obs_converter(obs_v, visual, obs_image)
+  File "/workspace/LLM-BC/llmbc/common/llfbench_util.py", line 39, in obs_converter
+    obs = torch.tensor(obs, dtype=torch.float32).unsqueeze(dim=0).to(device)
+KeyboardInterrupt

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/files/requirements.txt

@@ -0,0 +1,860 @@
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+llmbc==0.0.0
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+llmbc==0.0.0
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+llmbc==0.0.0
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+llmbc==0.0.0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.0.0
+inflect==7.3.1
+jaraco.collections==5.1.0
+jaraco.context==5.3.0
+jaraco.functools==4.0.1
+jaraco.text==3.12.1
+more-itertools==10.3.0
+packaging==24.2
+platformdirs==4.2.2
+tomli==2.0.1
+typeguard==4.3.0
+typing_extensions==4.12.2
+wheel==0.45.1
+zipp==3.19.2

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/files/wandb-metadata.json

@@ -0,0 +1,49 @@
+{
+  "os": "Linux-5.15.0-168-generic-x86_64-with-glibc2.39",
+  "python": "3.9.25",
+  "startedAt": "2026-04-10T09:16:57.612247Z",
+  "program": "/workspace/LLM-BC/./llmbc/workspace/train_bc_lowdim_workspace.py",
+  "codePath": "llmbc/workspace/train_bc_lowdim_workspace.py",
+  "git": {
+    "remote": "https://github.com/CHYang25/LLM-BC.git",
+    "commit": "2fa092f9eb4c2ad9435919c661a618071a583be5"
+  },
+  "email": "chris920325@gmail.com",
+  "root": "/workspace/LLM-BC/data/outputs/2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0",
+  "host": "c8ef52649900",
+  "username": "root",
+  "executable": "/venv/llm-bc/bin/python3",
+  "codePathLocal": "llmbc/workspace/train_bc_lowdim_workspace.py",
+  "cpu_count": 72,
+  "cpu_count_logical": 144,
+  "gpu": "NVIDIA GeForce RTX 4090",
+  "gpu_count": 2,
+  "disk": {
+    "/": {
+      "total": "545460846592",
+      "used": "206078439424"
+    }
+  },
+  "memory": {
+    "total": "810942898176"
+  },
+  "cpu": {
+    "count": 72,
+    "countLogical": 144
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA GeForce RTX 4090",
+      "memoryTotal": "25757220864",
+      "cudaCores": 16384,
+      "architecture": "Ada"
+    },
+    {
+      "name": "NVIDIA GeForce RTX 4090",
+      "memoryTotal": "25757220864",
+      "cudaCores": 16384,
+      "architecture": "Ada"
+    }
+  ],
+  "cudaVersion": "12.8"
+}

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_wandb":{"runtime":472},"global_step":197,"lr":0.009999978020164803,"_runtime":472.555626345,"_timestamp":1.7758126203738923e+09,"train_loss":0.36438101530075073,"epoch":0,"_step":197,"train_loss_bc":0.36438101530075073,"grad_norm":0.3752126395702362}

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/run-20260410_091657-ags56sz9/run-ags56sz9.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22f617861eb4d60e1dbaa3f0c287a22618d6896fbe0438b819f9419b852f7172
+size 425984

2026.04.10/09.16.06_train_bc_lowdim_maze2d-medium-v0/wandb/wandb-resume.json

@@ -0,0 +1 @@
+{"run_id": "ags56sz9"}

2026.04.10/09.26.57_train_llmbc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,165 @@
+name: train_llmbc_lowdim
+_target_: llmbc.workspace.train_llmbc_lowdim_workspace.TrainLLMBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+model_name: ${llm.name}
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+llm_orig_expert_feedback: true
+llm_do_sample: false
+policy:
+  _target_: llmbc.policy.llmbc_lowdim_policy.LLMBCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: tanh
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  llm_discriminator:
+    _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator
+    task_id: ${task_name}
+    llm_translator:
+      _target_: llmbc.translator.llm_translator.LLMTranslator
+      cfg: ${llm}
+      obs_dim: ${task.obs_dim}
+      action_dim: ${task.action_dim}
+      horizon: ${horizon}
+      n_obs_steps: ${n_obs_steps}
+      n_action_steps: ${n_action_steps}
+  loss_bc_weight: 1.0
+  loss_llm_weight: 0.01
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+  normalize_llm_loss: true
+dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.0001
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 801
+  gradient_accumulate_every: 8
+  grad_norm_clip: 0.5
+  rollout_every: 4
+  checkpoint_every: 4
+  val_every: 1
+  sample_every: 4
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training-mix
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0-mix(csp5).pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: true
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor
+llm:
+  name: HuggingFaceTB/SmolLM2-135M-Instruct
+  model_name: SmolLM2-135M-Instruct
+  config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig
+  causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM
+  use_quantization: false
+  use_joint_mlp_projector: true
+  llm_mode: ete-finetuned
+  finetune_mode: orig
+  checkpoint: data/outputs/2025.10.16/11.53.22_train_llm_lowdim_maze2d-medium-v0/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-maze2d-medium-v0/checkpoint-11850
+  max_length: 100
+  lora_config:
+    r: 32
+    lora_alpha: 64
+    lora_dropout: 0.05
+    bias: none
+    task_type: CAUSAL_LM
+  prompter:
+    _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter
+    use_joint_mlp_projector: true
+  hydra:
+    job:
+      override_dirname: ${model_name}
+    run:
+      dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name}

2026.04.10/09.26.57_train_llmbc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - training.seed=42
+  job:
+    name: train
+    chdir: null
+    override_dirname: training.seed=42
+    id: ???
+    num: ???
+    config_name: llmbc_maze2d-medium-v0_mix(csp5).yaml
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/config/mix_data
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.26.57_train_llmbc_lowdim_maze2d-medium-v0
+    choices:
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.26.57_train_llmbc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+- training.seed=42

2026.04.10/09.26.57_train_llmbc_lowdim_maze2d-medium-v0/train.log

@@ -0,0 +1,6 @@
+[2026-04-10 09:26:58,688][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
+[2026-04-10 09:26:58,688][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16.
+[2026-04-10 09:26:58,688][numexpr.utils][INFO] - NumExpr defaulting to 16 threads.
+[2026-04-10 09:27:02,106][datasets][INFO] - PyTorch version 2.2.2 available.
+[2026-04-10 09:27:02,106][datasets][INFO] - TensorFlow version 2.15.1 available.
+[2026-04-10 09:27:02,107][datasets][INFO] - JAX version 0.4.30 available.

2026.04.10/09.36.23_train_llmbc_lowdim_adroit-hand-relocate-v1/.hydra/config.yaml

@@ -0,0 +1,163 @@
+name: train_llmbc_lowdim
+_target_: llmbc.workspace.train_llmbc_lowdim_workspace.TrainLLMBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+model_name: ${llm.name}
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+llm_orig_expert_feedback: true
+llm_do_sample: false
+policy:
+  _target_: llmbc.policy.llmbc_lowdim_policy.LLMBCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: relu
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  llm_discriminator:
+    _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator
+    task_id: ${task_name}
+    llm_translator:
+      _target_: llmbc.translator.llm_translator.LLMTranslator
+      cfg: ${llm}
+      obs_dim: ${task.obs_dim}
+      action_dim: ${task.action_dim}
+      horizon: ${horizon}
+      n_obs_steps: ${n_obs_steps}
+      n_action_steps: ${n_action_steps}
+  loss_bc_weight: 1.0
+  loss_llm_weight: 0.01
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+  normalize_llm_loss: true
+dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.001
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: constant
+  lr_warmup_steps: 10
+  num_epochs: 1001
+  gradient_accumulate_every: 8
+  grad_norm_clip: 0.5
+  rollout_every: 5
+  checkpoint_every: 5
+  val_every: 1
+  sample_every: 5
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training-mix
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: adroit-hand-relocate-v1
+  obs_dim: 39
+  action_dim: 30
+  env_runner:
+    _target_: llmbc.env_runner.adroit_lowdim_runner.AdroitHandLowdimRunner
+    env_name: llf-adroit-adroit-hand-relocate-v1
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 150
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.adroit_lowdim_dataset.AdroitHandLowdimDataset
+    data_path: datasets/adroit-hand-relocate-v1-mix(csp100).pt
+    data_path2: datasets/adroit-hand-relocate-v1.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.1
+    dummy_normalizer: true
+  instructor:
+    _target_: llmbc.translator.instructor.adroit_instructor.adroit_hand_relocate_v1_instructor.AdroitHandRelocateV1Instructor
+llm:
+  name: HuggingFaceTB/SmolLM2-135M-Instruct
+  model_name: SmolLM2-135M-Instruct
+  config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig
+  causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM
+  use_quantization: false
+  use_joint_mlp_projector: true
+  llm_mode: ete-finetuned
+  finetune_mode: orig
+  checkpoint: data/outputs/2026.04.03/12.10.45_train_llm_lowdim_adroit-hand-relocate-v1/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-adroit-hand-relocate-v1/checkpoint-2340
+  max_length: 100
+  lora_config:
+    r: 32
+    lora_alpha: 64
+    lora_dropout: 0.05
+    bias: none
+    task_type: CAUSAL_LM
+  prompter:
+    _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter
+    use_joint_mlp_projector: true
+  hydra:
+    job:
+      override_dirname: ${model_name}
+    run:
+      dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name}

2026.04.10/09.36.23_train_llmbc_lowdim_adroit-hand-relocate-v1/.hydra/hydra.yaml

@@ -0,0 +1,154 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task: []
+  job:
+    name: train
+    chdir: null
+    override_dirname: ''
+    id: ???
+    num: ???
+    config_name: llmbc_adroit-hand-relocate-v1_mix(csp100).yaml
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/config/mix_data
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.36.23_train_llmbc_lowdim_adroit-hand-relocate-v1
+    choices:
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.36.23_train_llmbc_lowdim_adroit-hand-relocate-v1/.hydra/overrides.yaml

@@ -0,0 +1 @@
+[]

2026.04.10/09.36.23_train_llmbc_lowdim_adroit-hand-relocate-v1/train.log

@@ -0,0 +1,6 @@
+[2026-04-10 09:36:24,868][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
+[2026-04-10 09:36:24,868][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16.
+[2026-04-10 09:36:24,868][numexpr.utils][INFO] - NumExpr defaulting to 16 threads.
+[2026-04-10 09:36:28,715][datasets][INFO] - PyTorch version 2.2.2 available.
+[2026-04-10 09:36:28,716][datasets][INFO] - TensorFlow version 2.15.1 available.
+[2026-04-10 09:36:28,717][datasets][INFO] - JAX version 0.4.30 available.

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,165 @@
+name: train_llmbc_lowdim
+_target_: llmbc.workspace.train_llmbc_lowdim_workspace.TrainLLMBCLowdimWorkspace
+obs_dim: ${task.obs_dim}
+action_dim: ${task.action_dim}
+task_name: ${task.name}
+exp_name: default
+model_name: ${llm.name}
+horizon: 1
+n_obs_steps: 1
+n_action_steps: 1
+n_latency_steps: 0
+past_action_visible: false
+llm_orig_expert_feedback: true
+llm_do_sample: false
+policy:
+  _target_: llmbc.policy.llmbc_lowdim_policy.LLMBCLowdimPolicy
+  model:
+    _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+    input_size: ${eval:'${n_obs_steps}*${obs_dim}'}
+    hidden_size:
+    - 256
+    - 256
+    - 256
+    - 256
+    output_size: ${eval:'${n_action_steps}*${action_dim}'}
+    activation: tanh
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  llm_discriminator:
+    _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator
+    task_id: ${task_name}
+    llm_translator:
+      _target_: llmbc.translator.llm_translator.LLMTranslator
+      cfg: ${llm}
+      obs_dim: ${task.obs_dim}
+      action_dim: ${task.action_dim}
+      horizon: ${horizon}
+      n_obs_steps: ${n_obs_steps}
+      n_action_steps: ${n_action_steps}
+  loss_bc_weight: 1.0
+  loss_llm_weight: 0.01
+  horizon: ${horizon}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+  normalize_llm_loss: true
+dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 16
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.AdamW
+  lr: 0.0001
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda:0
+  seed: 42
+  debug: false
+  resume: false
+  lr_scheduler: cosine
+  lr_warmup_steps: 10
+  num_epochs: 801
+  gradient_accumulate_every: 8
+  grad_norm_clip: 0.5
+  rollout_every: 4
+  checkpoint_every: 4
+  val_every: 1
+  sample_every: 4
+  sample_max_batch: 128
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: ${task.name}-training-mix
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: test_success_rate
+    mode: max
+    k: 5
+    format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+task:
+  name: maze2d-medium-v0
+  obs_dim: 6
+  action_dim: 2
+  env_runner:
+    _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+    env_name: llf-pointmaze-maze2d-medium-v0
+    n_train: 10
+    n_test: 50
+    n_envs: 10
+    max_steps: 250
+    n_obs_steps: ${n_obs_steps}
+    n_action_steps: ${n_action_steps}
+    instruction_type: b
+    feedback_type:
+    - hp
+    - hn
+    - fp
+    visual: false
+    discount: 0.99
+  dataset:
+    _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+    data_path: datasets/maze2d-medium-v0-mix(csp5).pt
+    data_path2: datasets/maze2d-medium-v0.pt
+    horizon: ${horizon}
+    pad_before: ${eval:'${n_obs_steps}-1'}
+    pad_after: ${eval:'${n_action_steps}-1'}
+    obs_eef_target: true
+    use_manual_normalizer: false
+    val_ratio: 0.02
+    dummy_normalizer: true
+  instructor:
+    _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor
+llm:
+  name: HuggingFaceTB/SmolLM2-135M-Instruct
+  model_name: SmolLM2-135M-Instruct
+  config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig
+  causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM
+  use_quantization: false
+  use_joint_mlp_projector: true
+  llm_mode: ete-finetuned
+  finetune_mode: orig
+  checkpoint: data/outputs/2025.10.16/11.53.22_train_llm_lowdim_maze2d-medium-v0/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-maze2d-medium-v0/checkpoint-11850
+  max_length: 100
+  lora_config:
+    r: 32
+    lora_alpha: 64
+    lora_dropout: 0.05
+    bias: none
+    task_type: CAUSAL_LM
+  prompter:
+    _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter
+    use_joint_mlp_projector: true
+  hydra:
+    job:
+      override_dirname: ${model_name}
+    run:
+      dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name}

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/.hydra/hydra.yaml

@@ -0,0 +1,155 @@
+hydra:
+  run:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+  sweep:
+    dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - training.seed=42
+  job:
+    name: train
+    chdir: null
+    override_dirname: training.seed=42
+    id: ???
+    num: ???
+    config_name: llmbc_maze2d-medium-v0_mix(csp5).yaml
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/LLM-BC/config/mix_data
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/LLM-BC/data/outputs/2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0
+    choices:
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/.hydra/overrides.yaml

@@ -0,0 +1 @@
+- training.seed=42

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/train.log

@@ -0,0 +1,13 @@
+[2026-04-10 09:36:50,164][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
+[2026-04-10 09:36:50,164][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16.
+[2026-04-10 09:36:50,164][numexpr.utils][INFO] - NumExpr defaulting to 16 threads.
+[2026-04-10 09:36:53,596][datasets][INFO] - PyTorch version 2.2.2 available.
+[2026-04-10 09:36:53,597][datasets][INFO] - TensorFlow version 2.15.1 available.
+[2026-04-10 09:36:53,597][datasets][INFO] - JAX version 0.4.30 available.
+[2026-04-10 09:36:57,574][root][INFO] - running build_ext
+[2026-04-10 09:36:57,576][root][INFO] - building 'mujoco_py.cymj' extension
+[2026-04-10 09:36:57,576][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o -fopenmp -w
+[2026-04-10 09:37:39,277][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -fopenmp -w
+[2026-04-10 09:37:39,368][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -shared -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib -L/venv/llm-bc/lib -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -L/root/.mujoco/mujoco210/bin -Wl,--enable-new-dtags,-rpath,/root/.mujoco/mujoco210/bin -lmujoco210 -lglewosmesa -lOSMesa -lGL -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py/cymj.cpython-39-x86_64-linux-gnu.so -fopenmp
+[2026-04-10 09:37:41,989][absl][INFO] - MUJOCO_GL=osmesa, attempting to import specified OpenGL backend.
+[2026-04-10 09:37:41,996][absl][INFO] - MuJoCo library version is: 2.3.7

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/wandb/run-20260410_093742-2o6jqedz/files/config.yaml

@@ -0,0 +1,273 @@
+_target_:
+    value: llmbc.workspace.train_llmbc_lowdim_workspace.TrainLLMBCLowdimWorkspace
+_wandb:
+    value:
+        cli_version: 0.18.6
+        m: []
+        python_version: 3.9.25
+        t:
+            "1":
+                - 1
+                - 2
+                - 3
+                - 5
+                - 11
+                - 12
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 55
+                - 71
+                - 83
+                - 95
+                - 98
+                - 100
+                - 105
+            "2":
+                - 1
+                - 2
+                - 3
+                - 5
+                - 11
+                - 12
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 55
+                - 71
+                - 83
+                - 95
+                - 98
+                - 100
+                - 105
+            "3":
+                - 13
+                - 15
+                - 16
+                - 23
+                - 55
+                - 61
+            "4": 3.9.25
+            "5": 0.18.6
+            "6": 4.47.1
+            "8":
+                - 5
+            "12": 0.18.6
+            "13": linux-x86_64
+action_dim:
+    value: 2
+checkpoint:
+    value:
+        save_last_ckpt: true
+        save_last_snapshot: false
+        topk:
+            format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt
+            k: 5
+            mode: max
+            monitor_key: test_success_rate
+dataloader:
+    value:
+        batch_size: 16
+        num_workers: 0
+        persistent_workers: false
+        pin_memory: false
+        shuffle: true
+exp_name:
+    value: default
+horizon:
+    value: 1
+llm:
+    value:
+        causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM
+        checkpoint: data/outputs/2025.10.16/11.53.22_train_llm_lowdim_maze2d-medium-v0/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-maze2d-medium-v0/checkpoint-11850
+        config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig
+        finetune_mode: orig
+        hydra:
+            job:
+                override_dirname: HuggingFaceTB/SmolLM2-135M-Instruct
+            run:
+                dir: data/outputs/2026.04.10/09.36.48_HuggingFaceTB/SmolLM2-135M-Instruct
+        llm_mode: ete-finetuned
+        lora_config:
+            bias: none
+            lora_alpha: 64
+            lora_dropout: 0.05
+            r: 32
+            task_type: CAUSAL_LM
+        max_length: 100
+        model_name: SmolLM2-135M-Instruct
+        name: HuggingFaceTB/SmolLM2-135M-Instruct
+        prompter:
+            _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter
+            use_joint_mlp_projector: true
+        use_joint_mlp_projector: true
+        use_quantization: false
+llm_do_sample:
+    value: false
+llm_orig_expert_feedback:
+    value: true
+logging:
+    value:
+        group: null
+        id: null
+        mode: online
+        name: 2026.04.10-09.36.48_train_llmbc_lowdim_maze2d-medium-v0
+        project: maze2d-medium-v0-training-mix
+        resume: true
+        tags:
+            - train_llmbc_lowdim
+            - maze2d-medium-v0
+            - default
+model_name:
+    value: HuggingFaceTB/SmolLM2-135M-Instruct
+multi_run:
+    value:
+        run_dir: data/outputs/2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0
+        wandb_name_base: 2026.04.10-09.36.48_train_llmbc_lowdim_maze2d-medium-v0
+n_action_steps:
+    value: 1
+n_latency_steps:
+    value: 0
+n_obs_steps:
+    value: 1
+name:
+    value: train_llmbc_lowdim
+obs_dim:
+    value: 6
+optimizer:
+    value:
+        _target_: torch.optim.AdamW
+        betas:
+            - 0.95
+            - 0.999
+        eps: 1e-08
+        lr: 0.0001
+        weight_decay: 1e-06
+output_dir:
+    value: /workspace/LLM-BC/data/outputs/2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0
+past_action_visible:
+    value: false
+policy:
+    value:
+        _target_: llmbc.policy.llmbc_lowdim_policy.LLMBCLowdimPolicy
+        action_dim: 2
+        horizon: 1
+        llm_discriminator:
+            _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator
+            llm_translator:
+                _target_: llmbc.translator.llm_translator.LLMTranslator
+                action_dim: 2
+                cfg:
+                    causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM
+                    checkpoint: data/outputs/2025.10.16/11.53.22_train_llm_lowdim_maze2d-medium-v0/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-maze2d-medium-v0/checkpoint-11850
+                    config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig
+                    finetune_mode: orig
+                    hydra:
+                        job:
+                            override_dirname: HuggingFaceTB/SmolLM2-135M-Instruct
+                        run:
+                            dir: data/outputs/2026.04.10/09.36.48_HuggingFaceTB/SmolLM2-135M-Instruct
+                    llm_mode: ete-finetuned
+                    lora_config:
+                        bias: none
+                        lora_alpha: 64
+                        lora_dropout: 0.05
+                        r: 32
+                        task_type: CAUSAL_LM
+                    max_length: 100
+                    model_name: SmolLM2-135M-Instruct
+                    name: HuggingFaceTB/SmolLM2-135M-Instruct
+                    prompter:
+                        _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter
+                        use_joint_mlp_projector: true
+                    use_joint_mlp_projector: true
+                    use_quantization: false
+                horizon: 1
+                n_action_steps: 1
+                n_obs_steps: 1
+                obs_dim: 6
+            task_id: maze2d-medium-v0
+        loss_bc_weight: 1
+        loss_llm_weight: 0.01
+        model:
+            _target_: llmbc.model.policy.policy_mlp.PolicyMLP
+            activation: tanh
+            hidden_size:
+                - 256
+                - 256
+                - 256
+                - 256
+            input_size: 6
+            n_action_steps: 1
+            n_obs_steps: 1
+            output_size: 2
+        n_action_steps: 1
+        n_obs_steps: 1
+        normalize_llm_loss: true
+        obs_dim: 6
+task:
+    value:
+        action_dim: 2
+        dataset:
+            _target_: llmbc.dataset.pointmaze_lowdim_dataset.PointmazeLowdimDataset
+            data_path: datasets/maze2d-medium-v0-mix(csp5).pt
+            data_path2: datasets/maze2d-medium-v0.pt
+            dummy_normalizer: true
+            horizon: 1
+            obs_eef_target: true
+            pad_after: 0
+            pad_before: 0
+            use_manual_normalizer: false
+            val_ratio: 0.02
+        env_runner:
+            _target_: llmbc.env_runner.pointmaze_lowdim_runner.PointmazeLowdimRunner
+            discount: 0.99
+            env_name: llf-pointmaze-maze2d-medium-v0
+            feedback_type:
+                - hp
+                - hn
+                - fp
+            instruction_type: b
+            max_steps: 250
+            n_action_steps: 1
+            n_envs: 10
+            n_obs_steps: 1
+            n_test: 50
+            n_train: 10
+            visual: false
+        instructor:
+            _target_: llmbc.translator.instructor.pointmaze_instructor.maze2d_medium_v0_instructor.Maze2dMediumV0Instructor
+        name: maze2d-medium-v0
+        obs_dim: 6
+task_name:
+    value: maze2d-medium-v0
+training:
+    value:
+        checkpoint_every: 4
+        debug: false
+        device: cuda:0
+        grad_norm_clip: 0.5
+        gradient_accumulate_every: 8
+        lr_scheduler: cosine
+        lr_warmup_steps: 10
+        max_train_steps: null
+        max_val_steps: null
+        num_epochs: 801
+        resume: false
+        rollout_every: 4
+        sample_every: 4
+        sample_max_batch: 128
+        seed: 42
+        tqdm_interval_sec: 1
+        val_every: 1
+val_dataloader:
+    value:
+        batch_size: 16
+        num_workers: 0
+        persistent_workers: false
+        pin_memory: false
+        shuffle: true

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/wandb/run-20260410_093742-2o6jqedz/files/output.log

@@ -0,0 +1,52 @@
+Traceback (most recent call last):                                                                                                                                                 
+  File "/workspace/LLM-BC/./train.py", line 35, in <module>
+    main()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/main.py", line 90, in decorated_main
+    _run_hydra(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 389, in _run_hydra
+    _run_app(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 452, in _run_app
+    run_and_report(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 213, in run_and_report
+    return func()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 453, in <lambda>
+    lambda: hydra.run(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/hydra.py", line 119, in run
+    ret = run_job(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/core/utils.py", line 186, in run_job
+    ret.return_value = task_function(task_cfg)
+  File "/workspace/LLM-BC/./train.py", line 32, in main
+    workspace.run()
+  File "/workspace/LLM-BC/llmbc/workspace/train_llmbc_lowdim_workspace.py", line 186, in run
+    loss.backward()
+  File "/venv/llm-bc/lib/python3.9/site-packages/torch/_tensor.py", line 522, in backward
+    torch.autograd.backward(
+  File "/venv/llm-bc/lib/python3.9/site-packages/torch/autograd/__init__.py", line 266, in backward
+    Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+KeyboardInterrupt
+Traceback (most recent call last):
+  File "/workspace/LLM-BC/./train.py", line 35, in <module>
+    main()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/main.py", line 90, in decorated_main
+    _run_hydra(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 389, in _run_hydra
+    _run_app(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 452, in _run_app
+    run_and_report(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 213, in run_and_report
+    return func()
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/utils.py", line 453, in <lambda>
+    lambda: hydra.run(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/_internal/hydra.py", line 119, in run
+    ret = run_job(
+  File "/venv/llm-bc/lib/python3.9/site-packages/hydra/core/utils.py", line 186, in run_job
+    ret.return_value = task_function(task_cfg)
+  File "/workspace/LLM-BC/./train.py", line 32, in main
+    workspace.run()
+  File "/workspace/LLM-BC/llmbc/workspace/train_llmbc_lowdim_workspace.py", line 186, in run
+    loss.backward()
+  File "/venv/llm-bc/lib/python3.9/site-packages/torch/_tensor.py", line 522, in backward
+    torch.autograd.backward(
+  File "/venv/llm-bc/lib/python3.9/site-packages/torch/autograd/__init__.py", line 266, in backward
+    Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+KeyboardInterrupt

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/wandb/run-20260410_093742-2o6jqedz/files/requirements.txt

@@ -0,0 +1,861 @@
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+llmbc==0.0.0
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+llmbc==0.0.0
+llmbc==0.0.0
+llmbc==0.0.0
+aiohappyeyeballs==2.6.1
+antlr4-python3-runtime==4.9.3
+asciitree==0.3.3
+async-timeout==5.0.1
+attrs==25.4.0
+certifi==2025.10.5
+charset-normalizer==3.4.4
+colorama==0.4.6
+dill==0.3.8
+docstring-parser==0.16
+eval_type_backport==0.2.2
+fasteners==0.20
+filelock==3.17.0
+frozenlist==1.8.0
+fsspec==2024.3.1
+gmpy2==2.2.1
+hf-xet==1.1.8
+idna==3.11
+llvmlite==0.39.1
+MarkupSafe==3.0.2
+mdurl==0.1.2
+mkl-service==2.4.0
+mpmath==1.3.0
+msgpack==1.1.1
+networkx==3.2.1
+packaging==25.0
+propcache==0.3.1
+psutil==7.0.0
+pycparser==2.23
+Pygments==2.19.1
+PyOpenGL==3.1.10
+PySocks==1.7.1
+tzdata==2025.3
+xxhash==3.5.0
+pytz==2025.2
+PyYAML==6.0.2
+regex==2025.9.1
+safetensors==0.5.3
+setuptools==80.9.0
+shtab==1.7.2
+six==1.17.0
+tqdm==4.67.1
+tensorboard==2.15.2
+uvloop==0.22.1
+wheel==0.45.1
+zipp==3.21.0
+tensorflow-estimator==2.15.0
+tensorboard-data-server==0.7.2
+google-auth-oauthlib==1.3.1
+tensorflow==2.15.1
+Jinja2==3.1.6
+markdown-it-py==2.2.0
+multiprocess==0.70.15
+numpy==1.23.5
+pillow==11.3.0
+pip==23.3.2
+python-dateutil==2.9.0.post0
+aiosignal==1.4.0
+brotlicffi==1.0.9.2
+hydra-core==1.2.0
+multidict==6.7.0
+pycares==4.10.0
+rich==14.2.0
+sympy==1.14.0
+typeguard==4.4.4
+aiodns==3.5.0
+urllib3==2.5.0
+yarl==1.18.0
+aiohttp==3.12.15
+pyarrow==21.0.0
+pyarrow-hotfix==0.7
+tokenizers==0.21.0
+av==10.0.0
+einops==0.4.1
+Bottleneck==1.4.2
+mkl_fft==1.3.11
+mkl_random==1.2.8
+numba==0.56.4
+numcodecs==0.12.1
+numexpr==2.10.1
+pandas==2.3.3
+zarr==2.12.0
+datasets==2.19.0
+transformers==4.47.1
+accelerate==1.0.1
+peft==0.14.0
+trl==0.11.4
+torch==2.2.2
+tabulate==0.9.0
+sentencepiece==0.2.0
+pyperclip==1.11.0
+PyOpenGL-accelerate==3.1.10
+pydub==0.25.1
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+nvidia-ml-py==13.595.45
+hjson==3.1.0
+gym-notices==0.1.0
+glfw==2.0.0
+Farama-Notifications==0.0.4
+websockets==15.0.1
+wcwidth==0.6.0
+uc-micro-py==1.0.3
+typing-inspection==0.4.2
+triton==2.2.0
+trimesh==4.11.5
+transforms3d==0.4.2
+traitlets==5.14.3
+threadpoolctl==3.6.0
+tenacity==9.1.2
+sniffio==1.3.1
+smmap==5.0.3
+setproctitle==1.3.7
+sentry-sdk==2.57.0
+semantic-version==2.10.0
+scipy==1.13.1
+rpds-py==0.27.1
+python-multipart==0.0.20
+pyparsing==3.3.2
+pynvml==13.0.1
+pydantic_core==2.41.5
+pycryptodomex==3.23.0
+pyasn1==0.6.3
+grpcio==1.80.0
+platformdirs==4.4.0
+pexpect==4.9.0
+parso==0.8.6
+orjson==3.11.5
+opencv-python==4.11.0.86
+stack-data==0.6.3
+omegaconf==2.2.1
+nvidia-nvtx-cu12==12.1.105
+nvidia-nvjitlink-cu12==12.9.86
+nvidia-nccl-cu12==2.19.3
+nvidia-curand-cu12==10.3.2.106
+nvidia-cufft-cu12==11.0.2.54
+nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cublas-cu12==12.1.3.1
+ninja==1.13.0
+narwhals==2.19.0
+lxml==6.0.3
+kiwisolver==1.4.7
+joblib==1.5.3
+jiter==0.13.0
+imageio-ffmpeg==0.6.0
+ImageIO==2.37.2
+h5py==3.14.0
+h11==0.16.0
+fonttools==4.60.2
+ffmpy==1.0.0
+fast_kinematics==0.2.2
+executing==2.2.1
+exceptiongroup==1.3.1
+docker-pycreds==0.4.0
+distro==1.9.0
+decorator==5.2.1
+dacite==1.9.2
+cycler==0.12.1
+contourpy==1.3.0
+cloudpickle==3.1.2
+click==8.1.8
+asttokens==3.0.1
+annotated-types==0.7.0
+annotated-doc==0.0.4
+aiofiles==25.1.0
+absl-py==2.3.1
+uvicorn==0.39.0
+tiktoken==0.8.0
+scikit-learn==1.6.1
+sapien==3.0.0b1
+referencing==0.36.2
+pydantic==2.12.5
+pyasn1_modules==0.4.2
+prompt_toolkit==3.0.52
+nvidia-cusparse-cu12==12.1.0.106
+nvidia-cudnn-cu12==8.9.2.26
+nltk==3.9.2
+mujoco==2.3.7
+mdit-py-plugins==0.3.3
+matplotlib-inline==0.2.1
+matplotlib==3.7.5
+linkify-it-py==2.0.3
+jedi==0.19.2
+huggingface-hub==0.26.2
+httpcore==1.0.9
+gymnasium==0.29.1
+google-pasta==0.2.0
+gitdb==4.0.12
+cryptography==46.0.7
+blobfile==3.0.0
+anyio==4.12.1
+tyro==0.9.1
+toppra==0.6.6
+starlette==0.49.3
+seaborn==0.13.2
+rouge_score==0.1.2
+pynndescent==0.6.0
+nvidia-cusolver-cu12==11.4.5.107
+jsonschema-specifications==2025.9.1
+ipython==8.18.1
+httpx==0.28.1
+gymnasium-robotics==1.2.0
+google-auth==2.49.2
+GitPython==3.1.46
+diffusers==0.31.0
+wandb==0.18.6
+umap-learn==0.5.9.post2
+requests-oauthlib==2.0.0
+openai==2.8.1
+mplib==0.1.1
+jsonschema==4.25.1
+gradio_client==0.2.9
+google-genai==1.47.0
+fastapi==0.128.8
+torchvision==0.17.2
+torchaudio==2.2.2
+sentence-transformers==3.2.1
+pytorch-seed==0.2.0
+deepspeed==0.16.1
+bitsandbytes==0.45.0
+altair==6.0.0
+timm==1.0.22
+gradio==3.36.1
+evaluate==0.4.3
+arm_pytorch_utilities==0.5.0
+tensorflow-io-gcs-filesystem==0.37.1
+Markdown==3.9
+pybullet==3.2.6
+wrapt==1.14.2
+tensorflow-probability==0.23.0
+parse==1.19.1
+dm-tree==0.1.8
+termcolor==3.1.0
+requests==2.32.0
+cffi==2.0.0
+opt_einsum==3.4.0
+labmaze==1.0.6
+typing_extensions==4.15.0
+importlib-resources==5.13.0
+importlib-metadata==5.2.0
+FLAML==2.3.6
+dm-env==1.6
+diskcache==5.6.3
+Cython==0.29.37
+mujoco-py==2.1.2.14
+jaxlib==0.4.30
+dm-control==1.0.14
+cmudict==1.0.13
+syllables==1.0.9
+metaworld==2.0.0
+jax==0.4.30
+highway-env==1.9.1
+gym_bandits==0.0.2
+d4rl==1.1
+pyautogen==0.1.0
+protobuf==4.25.9
+tf-agents==0.19.0
+oauthlib==3.3.1
+pytorch-kinematics==0.7.5
+pygame==2.6.1
+gin-config==0.5.0
+mani_skill==3.0.0b20
+tifffile==2024.8.30
+Shapely==1.8.4
+PyWavelets==1.6.0
+Werkzeug==3.1.8
+astunparse==1.6.3
+scikit-image==0.19.3
+pymunk==6.2.1
+ml-dtypes==0.3.2
+gast==0.7.0
+stable-baselines3==2.2.1
+keras==2.15.0
+llfbench==0.1.0
+libclang==18.1.1
+flatbuffers==25.12.19
+gym==0.23.1
+llmbc==0.0.0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.0.0
+inflect==7.3.1
+jaraco.collections==5.1.0
+jaraco.context==5.3.0
+jaraco.functools==4.0.1
+jaraco.text==3.12.1
+more-itertools==10.3.0
+packaging==24.2
+platformdirs==4.2.2
+tomli==2.0.1
+typeguard==4.3.0
+typing_extensions==4.12.2
+wheel==0.45.1
+zipp==3.19.2

2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0/wandb/run-20260410_093742-2o6jqedz/files/wandb-metadata.json

@@ -0,0 +1,56 @@
+{
+  "os": "Linux-5.15.0-168-generic-x86_64-with-glibc2.39",
+  "python": "3.9.25",
+  "startedAt": "2026-04-10T09:37:42.923208Z",
+  "args": [
+    "--config-path",
+    "config/mix_data",
+    "--config-name",
+    "llmbc_maze2d-medium-v0_mix(csp5).yaml",
+    "training.seed=42"
+  ],
+  "program": "/workspace/LLM-BC/./train.py",
+  "codePath": "train.py",
+  "git": {
+    "remote": "https://github.com/CHYang25/LLM-BC.git",
+    "commit": "2fa092f9eb4c2ad9435919c661a618071a583be5"
+  },
+  "email": "chris920325@gmail.com",
+  "root": "/workspace/LLM-BC/data/outputs/2026.04.10/09.36.48_train_llmbc_lowdim_maze2d-medium-v0",
+  "host": "c8ef52649900",
+  "username": "root",
+  "executable": "/venv/llm-bc/bin/python3",
+  "codePathLocal": "train.py",
+  "cpu_count": 72,
+  "cpu_count_logical": 144,
+  "gpu": "NVIDIA GeForce RTX 4090",
+  "gpu_count": 2,
+  "disk": {
+    "/": {
+      "total": "545460846592",
+      "used": "335971295232"
+    }
+  },
+  "memory": {
+    "total": "810942898176"
+  },
+  "cpu": {
+    "count": 72,
+    "countLogical": 144
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA GeForce RTX 4090",
+      "memoryTotal": "25757220864",
+      "cudaCores": 16384,
+      "architecture": "Ada"
+    },
+    {
+      "name": "NVIDIA GeForce RTX 4090",
+      "memoryTotal": "25757220864",
+      "cudaCores": 16384,
+      "architecture": "Ada"
+    }
+  ],
+  "cudaVersion": "12.8"
+}