Upload folder using huggingface_hub

.gitattributes

@@ -263,3 +263,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 2026.07.26/02.56.53_train_rbc_reward_model_adroit-hand-relocate-v1/wandb/run-20260726_025657-j968ki3f/run-j968ki3f.wandb filter=lfs diff=lfs merge=lfs -text
 2026.07.25/15.13.25_train_rbc_reward_model_maze2d-medium-v0/logs.json.txt filter=lfs diff=lfs merge=lfs -text
 2026.07.25/15.13.25_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151328-qq2i95ep/run-qq2i95ep.wandb filter=lfs diff=lfs merge=lfs -text
+2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/logs.json.txt filter=lfs diff=lfs merge=lfs -text
+2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/run-i12soqic.wandb filter=lfs diff=lfs merge=lfs -text

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/.hydra/config.yaml

@@ -0,0 +1,114 @@
+name: train_rbc_reward_model
+_target_: llmbc.workspace.train_rbc_reward_workspace.TrainRBCRewardWorkspace
+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
+reward_model:
+  _target_: llmbc.policy.rbc_lowdim_policy.RBCLowdimRewardModel
+  input_size: ${eval:'${n_obs_steps}*${obs_dim}+${n_action_steps}*${action_dim}'}
+  hidden_size:
+  - 256
+  - 256
+  - 256
+  obs_dim: ${obs_dim}
+  action_dim: ${action_dim}
+  n_obs_steps: ${n_obs_steps}
+  n_action_steps: ${n_action_steps}
+dataloader:
+  batch_size: 256
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+val_dataloader:
+  batch_size: 256
+  num_workers: 0
+  shuffle: true
+  pin_memory: false
+  persistent_workers: false
+optimizer:
+  _target_: torch.optim.Adam
+  lr: 0.0001
+  betas:
+  - 0.95
+  - 0.999
+  eps: 1.0e-08
+  weight_decay: 1.0e-06
+training:
+  device: cuda
+  seed: 42
+  debug: false
+  resume: true
+  lr_scheduler: cosine
+  lr_warmup_steps: 100
+  num_epochs: 2000
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  checkpoint_every: 50
+  val_every: 1
+  sample_every: 50
+  sample_max_batch: 1000
+  max_train_steps: null
+  max_val_steps: null
+  tqdm_interval_sec: 1.0
+logging:
+  project: rbc_module_training
+  resume: true
+  mode: online
+  name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name}
+  tags:
+  - ${name}
+  - ${task_name}
+  - ${exp_name}
+  id: null
+  group: null
+checkpoint:
+  topk:
+    monitor_key: val_loss
+    mode: min
+    k: 5
+    format_str: epoch={epoch:04d}-val_loss={val_loss:.3f}.ckpt
+  save_last_ckpt: true
+  save_last_snapshot: false
+multi_run:
+  run_dir: data/outputs/${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-general-freward.pt
+    data_path2: datasets/maze2d-medium-v0-freward.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

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/.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: rbc_reward-freward_maze2d-medium-v0.yaml
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.2.0
+    version_base: '1.2'
+    cwd: /tmp2/chyang/workspace/LLM-BC
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /tmp2/chyang/workspace/LLM-BC/config/main_table
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_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

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/.hydra/overrides.yaml

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

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/epoch=0100-val_loss=1.066.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caeea24572ddf6d34de8d8a58dcfce085ee11e8e7953cb047b6d7dbf3b2deb2a
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/epoch=0200-val_loss=1.059.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f07347393b0581ca62234c41641e586e79c9f0f87118898dce8a57c55ac0748
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/epoch=0250-val_loss=1.068.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f6bbdd7b3bed8a046fbd1d03f68c3fd147870e0d70dca2cb21eeba93742ddc0
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/epoch=0300-val_loss=1.007.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aac4d27094afe1fa823703f680717582d714ddc72c22616e64f66ef2a61ba103
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/epoch=0350-val_loss=1.051.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16e43c7b979be7298551969aba92795722aa6e17720197a208e11b481bfe761f
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/checkpoints/latest.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2436652b17cf22280c2cad10f66a7044c7151ac6dddc6afeb18bcfc3e3188bf6
+size 1634992

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/logs.json.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4713aee37fe9829037edc70eef2eca2881f57ea3d019218cea7cfb8d1b7e89b1
+size 158418131

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/train.log

@@ -0,0 +1,2 @@
+[2026-07-25 15:13:00,637][numexpr.utils][INFO] - Note: NumExpr detected 20 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16.
+[2026-07-25 15:13:00,638][numexpr.utils][INFO] - NumExpr defaulting to 16 threads.

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2026-07-25T15:13:03.725013663+08:00","level":"INFO","msg":"using version","core version":"0.18.6"}
+{"time":"2026-07-25T15:13:03.725038255+08:00","level":"INFO","msg":"created symlink","path":"/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-core.log"}
+{"time":"2026-07-25T15:13:03.833992104+08:00","level":"INFO","msg":"created new stream","id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834045529+08:00","level":"INFO","msg":"stream: started","id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834083091+08:00","level":"INFO","msg":"sender: started","stream_id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834079187+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T15:13:03.834084014+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T15:13:04.597575202+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-07-25T17:07:31.921426521+08:00","level":"INFO","msg":"stream: closing","id":"i12soqic"}
+{"time":"2026-07-25T17:07:31.921480084+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-07-25T17:07:31.922308379+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-07-25T17:07:35.75383166+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-07-25T17:07:36.580949644+08:00","level":"INFO","msg":"handler: closed","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T17:07:36.581040319+08:00","level":"INFO","msg":"sender: closed","stream_id":"i12soqic"}
+{"time":"2026-07-25T17:07:36.581018256+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T17:07:36.581170711+08:00","level":"INFO","msg":"stream: closed","id":"i12soqic"}

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/debug.log

@@ -0,0 +1,27 @@
+2026-07-25 15:13:03,717 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Current SDK version is 0.18.6
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Configure stats pid to 707073
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from /home/chyang/.config/wandb/settings
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from /tmp2/chyang/workspace/LLM-BC/wandb/settings
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': 'online', '_disable_service': None}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Inferring run settings from compute environment: {'program_relpath': 'train.py', 'program_abspath': '/tmp2/chyang/workspace/LLM-BC/train.py', 'program': '/tmp2/chyang/workspace/LLM-BC/./train.py'}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:_log_setup():533] Logging user logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug.log
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:_log_setup():534] Logging internal logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-internal.log
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:init():619] calling init triggers
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():626] wandb.init called with sweep_config: {}
+config: {'name': 'train_rbc_reward_model', '_target_': 'llmbc.workspace.train_rbc_reward_workspace.TrainRBCRewardWorkspace', 'obs_dim': 6, 'action_dim': 2, 'task_name': 'maze2d-medium-v0', 'exp_name': 'default', 'horizon': 1, 'n_obs_steps': 1, 'n_action_steps': 1, 'n_latency_steps': 0, 'past_action_visible': False, 'reward_model': {'_target_': 'llmbc.policy.rbc_lowdim_policy.RBCLowdimRewardModel', 'input_size': 8, 'hidden_size': [256, 256, 256], 'obs_dim': 6, 'action_dim': 2, 'n_obs_steps': 1, 'n_action_steps': 1}, 'dataloader': {'batch_size': 256, 'num_workers': 0, 'shuffle': True, 'pin_memory': False, 'persistent_workers': False}, 'val_dataloader': {'batch_size': 256, 'num_workers': 0, 'shuffle': True, 'pin_memory': False, 'persistent_workers': False}, 'optimizer': {'_target_': 'torch.optim.Adam', 'lr': 0.0001, 'betas': [0.95, 0.999], 'eps': 1e-08, 'weight_decay': 1e-06}, 'training': {'device': 'cuda', 'seed': 42, 'debug': False, 'resume': True, 'lr_scheduler': 'cosine', 'lr_warmup_steps': 100, 'num_epochs': 2000, 'gradient_accumulate_every': 1, 'grad_norm_clip': 0.5, 'checkpoint_every': 50, 'val_every': 1, 'sample_every': 50, 'sample_max_batch': 1000, 'max_train_steps': None, 'max_val_steps': None, 'tqdm_interval_sec': 1.0}, 'logging': {'project': 'rbc_module_training', 'resume': True, 'mode': 'online', 'name': '2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0', 'tags': ['train_rbc_reward_model', 'maze2d-medium-v0', 'default'], 'id': None, 'group': None}, 'checkpoint': {'topk': {'monitor_key': 'val_loss', 'mode': 'min', 'k': 5, 'format_str': 'epoch={epoch:04d}-val_loss={val_loss:.3f}.ckpt'}, 'save_last_ckpt': True, 'save_last_snapshot': False}, 'multi_run': {'run_dir': 'data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0', 'wandb_name_base': '2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0'}, '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': 1, 'n_action_steps': 1, '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-general-freward.pt', 'data_path2': 'datasets/maze2d-medium-v0-freward.pt', 'horizon': 1, 'pad_before': 0, 'pad_after': 0, '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'}}}
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():669] starting backend
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():673] sending inform_init request
+2026-07-25 15:13:03,721 INFO    MainThread:707073 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-07-25 15:13:03,722 INFO    MainThread:707073 [wandb_init.py:init():686] backend started and connected
+2026-07-25 15:13:03,729 INFO    MainThread:707073 [wandb_init.py:init():781] updated telemetry
+2026-07-25 15:13:03,762 INFO    MainThread:707073 [wandb_init.py:init():814] communicating run to backend with 90.0 second timeout
+2026-07-25 15:13:04,591 INFO    MainThread:707073 [wandb_init.py:init():867] starting run threads in backend
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_console_start():2451] atexit reg
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2299] redirect: wrap_raw
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2364] Wrapping output streams.
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2389] Redirects installed.
+2026-07-25 15:13:04,777 INFO    MainThread:707073 [wandb_init.py:init():911] run started, returning control to user process
+2026-07-25 15:13:04,777 INFO    MainThread:707073 [wandb_run.py:_config_callback():1389] config_cb None None {'output_dir': '/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0'}
+2026-07-25 17:07:31,921 WARNING MsgRouterThr:707073 [router.py:message_loop():75] message_loop has been closed

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/files/config.yaml

@@ -0,0 +1,170 @@
+_target_:
+    value: llmbc.workspace.train_rbc_reward_workspace.TrainRBCRewardWorkspace
+_wandb:
+    value:
+        cli_version: 0.18.6
+        m: []
+        python_version: 3.9.21
+        t:
+            "1":
+                - 1
+                - 11
+                - 49
+                - 50
+                - 55
+                - 83
+            "2":
+                - 1
+                - 11
+                - 49
+                - 50
+                - 55
+                - 83
+            "3":
+                - 13
+                - 15
+                - 16
+                - 23
+                - 55
+                - 61
+            "4": 3.9.21
+            "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}-val_loss={val_loss:.3f}.ckpt
+            k: 5
+            mode: min
+            monitor_key: val_loss
+dataloader:
+    value:
+        batch_size: 256
+        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: 2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0
+        project: rbc_module_training
+        resume: true
+        tags:
+            - train_rbc_reward_model
+            - maze2d-medium-v0
+            - default
+multi_run:
+    value:
+        run_dir: data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0
+        wandb_name_base: 2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0
+n_action_steps:
+    value: 1
+n_latency_steps:
+    value: 0
+n_obs_steps:
+    value: 1
+name:
+    value: train_rbc_reward_model
+obs_dim:
+    value: 6
+optimizer:
+    value:
+        _target_: torch.optim.Adam
+        betas:
+            - 0.95
+            - 0.999
+        eps: 1e-08
+        lr: 0.0001
+        weight_decay: 1e-06
+output_dir:
+    value: /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0
+past_action_visible:
+    value: false
+reward_model:
+    value:
+        _target_: llmbc.policy.rbc_lowdim_policy.RBCLowdimRewardModel
+        action_dim: 2
+        hidden_size:
+            - 256
+            - 256
+            - 256
+        input_size: 8
+        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-general-freward.pt
+            data_path2: datasets/maze2d-medium-v0-freward.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: 50
+        debug: false
+        device: cuda
+        grad_norm_clip: 0.5
+        gradient_accumulate_every: 1
+        lr_scheduler: cosine
+        lr_warmup_steps: 100
+        max_train_steps: null
+        max_val_steps: null
+        num_epochs: 2000
+        resume: true
+        sample_every: 50
+        sample_max_batch: 1000
+        seed: 42
+        tqdm_interval_sec: 1
+        val_every: 1
+val_dataloader:
+    value:
+        batch_size: 256
+        num_workers: 0
+        persistent_workers: false
+        pin_memory: false
+        shuffle: true

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/files/output.log

@@ -0,0 +1 @@
+                                                                                                                                                                                  

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/files/wandb-metadata.json

@@ -0,0 +1,55 @@
+{
+  "os": "Linux-5.15.0-139-generic-x86_64-with-glibc2.35",
+  "python": "3.9.21",
+  "startedAt": "2026-07-25T07:13:03.722547Z",
+  "args": [
+    "--config-path",
+    "config/main_table",
+    "--config-name",
+    "rbc_reward-freward_maze2d-medium-v0.yaml"
+  ],
+  "program": "/tmp2/chyang/workspace/LLM-BC/./train.py",
+  "codePath": "train.py",
+  "git": {
+    "remote": "https://github.com/CHYang25/LLM-BC.git",
+    "commit": "d38739f39371e5f7668c54610f490804f625fdb7"
+  },
+  "email": "chris920325@gmail.com",
+  "root": "/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0",
+  "host": "user",
+  "username": "chyang",
+  "executable": "/home/chyang/miniconda3/envs/llm-bc/bin/python3",
+  "codePathLocal": "train.py",
+  "cpu_count": 10,
+  "cpu_count_logical": 20,
+  "gpu": "NVIDIA GeForce RTX 3090",
+  "gpu_count": 2,
+  "disk": {
+    "/": {
+      "total": "925842214912",
+      "used": "873665310720"
+    }
+  },
+  "memory": {
+    "total": "134733234176"
+  },
+  "cpu": {
+    "count": 10,
+    "countLogical": 20
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA GeForce RTX 3090",
+      "memoryTotal": "25769803776",
+      "cudaCores": 10496,
+      "architecture": "Ampere"
+    },
+    {
+      "name": "NVIDIA GeForce RTX 3080",
+      "memoryTotal": "10737418240",
+      "cudaCores": 8704,
+      "architecture": "Ampere"
+    }
+  ],
+  "cudaVersion": "12.9"
+}

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_step":1175999,"train_batch_mse_error":0.11720578372478485,"_timestamp":1.7849704519196866e+09,"grad_norm":18.49838638305664,"_runtime":6868.198914503,"global_step":1175999,"_wandb":{"runtime":6868},"train_loss":0.14235418398870903,"val_loss":1.2379237413406372,"epoch":1999,"lr":0}

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-core.log

@@ -0,0 +1,14 @@
+{"time":"2026-07-25T15:13:02.876245663+08:00","level":"INFO","msg":"started logging, with flags","port-filename":"/tmp/tmpk87vbe4a/port-707073.txt","pid":707073,"debug":false,"disable-analytics":false}
+{"time":"2026-07-25T15:13:02.876271191+08:00","level":"INFO","msg":"FeatureState","shutdownOnParentExitEnabled":false}
+{"time":"2026-07-25T15:13:02.878995638+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":43513,"Zone":""}}
+{"time":"2026-07-25T15:13:02.8790294+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":707073}
+{"time":"2026-07-25T15:13:03.071000572+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T15:13:03.724769963+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"i12soqic","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T15:13:03.834056109+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"i12soqic","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:31.921332314+08:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:31.921421494+08:00","level":"INFO","msg":"connection: Close: initiating connection closure","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:31.92144314+08:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-07-25T17:07:31.92154435+08:00","level":"INFO","msg":"connection: Close: connection successfully closed","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:36.581270464+08:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:36.581314717+08:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"127.0.0.1:43038"}
+{"time":"2026-07-25T17:07:36.581340085+08:00","level":"INFO","msg":"server is closed"}

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2026-07-25T15:13:03.725013663+08:00","level":"INFO","msg":"using version","core version":"0.18.6"}
+{"time":"2026-07-25T15:13:03.725038255+08:00","level":"INFO","msg":"created symlink","path":"/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-core.log"}
+{"time":"2026-07-25T15:13:03.833992104+08:00","level":"INFO","msg":"created new stream","id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834045529+08:00","level":"INFO","msg":"stream: started","id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834083091+08:00","level":"INFO","msg":"sender: started","stream_id":"i12soqic"}
+{"time":"2026-07-25T15:13:03.834079187+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T15:13:03.834084014+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T15:13:04.597575202+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-07-25T17:07:31.921426521+08:00","level":"INFO","msg":"stream: closing","id":"i12soqic"}
+{"time":"2026-07-25T17:07:31.921480084+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-07-25T17:07:31.922308379+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-07-25T17:07:35.75383166+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-07-25T17:07:36.580949644+08:00","level":"INFO","msg":"handler: closed","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T17:07:36.581040319+08:00","level":"INFO","msg":"sender: closed","stream_id":"i12soqic"}
+{"time":"2026-07-25T17:07:36.581018256+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":{"value":"i12soqic"}}
+{"time":"2026-07-25T17:07:36.581170711+08:00","level":"INFO","msg":"stream: closed","id":"i12soqic"}

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug.log

@@ -0,0 +1,27 @@
+2026-07-25 15:13:03,717 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Current SDK version is 0.18.6
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Configure stats pid to 707073
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from /home/chyang/.config/wandb/settings
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from /tmp2/chyang/workspace/LLM-BC/wandb/settings
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': 'online', '_disable_service': None}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Inferring run settings from compute environment: {'program_relpath': 'train.py', 'program_abspath': '/tmp2/chyang/workspace/LLM-BC/train.py', 'program': '/tmp2/chyang/workspace/LLM-BC/./train.py'}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:_log_setup():533] Logging user logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug.log
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:_log_setup():534] Logging internal logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/logs/debug-internal.log
+2026-07-25 15:13:03,718 INFO    MainThread:707073 [wandb_init.py:init():619] calling init triggers
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():626] wandb.init called with sweep_config: {}
+config: {'name': 'train_rbc_reward_model', '_target_': 'llmbc.workspace.train_rbc_reward_workspace.TrainRBCRewardWorkspace', 'obs_dim': 6, 'action_dim': 2, 'task_name': 'maze2d-medium-v0', 'exp_name': 'default', 'horizon': 1, 'n_obs_steps': 1, 'n_action_steps': 1, 'n_latency_steps': 0, 'past_action_visible': False, 'reward_model': {'_target_': 'llmbc.policy.rbc_lowdim_policy.RBCLowdimRewardModel', 'input_size': 8, 'hidden_size': [256, 256, 256], 'obs_dim': 6, 'action_dim': 2, 'n_obs_steps': 1, 'n_action_steps': 1}, 'dataloader': {'batch_size': 256, 'num_workers': 0, 'shuffle': True, 'pin_memory': False, 'persistent_workers': False}, 'val_dataloader': {'batch_size': 256, 'num_workers': 0, 'shuffle': True, 'pin_memory': False, 'persistent_workers': False}, 'optimizer': {'_target_': 'torch.optim.Adam', 'lr': 0.0001, 'betas': [0.95, 0.999], 'eps': 1e-08, 'weight_decay': 1e-06}, 'training': {'device': 'cuda', 'seed': 42, 'debug': False, 'resume': True, 'lr_scheduler': 'cosine', 'lr_warmup_steps': 100, 'num_epochs': 2000, 'gradient_accumulate_every': 1, 'grad_norm_clip': 0.5, 'checkpoint_every': 50, 'val_every': 1, 'sample_every': 50, 'sample_max_batch': 1000, 'max_train_steps': None, 'max_val_steps': None, 'tqdm_interval_sec': 1.0}, 'logging': {'project': 'rbc_module_training', 'resume': True, 'mode': 'online', 'name': '2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0', 'tags': ['train_rbc_reward_model', 'maze2d-medium-v0', 'default'], 'id': None, 'group': None}, 'checkpoint': {'topk': {'monitor_key': 'val_loss', 'mode': 'min', 'k': 5, 'format_str': 'epoch={epoch:04d}-val_loss={val_loss:.3f}.ckpt'}, 'save_last_ckpt': True, 'save_last_snapshot': False}, 'multi_run': {'run_dir': 'data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0', 'wandb_name_base': '2026.07.25-15.12.59_train_rbc_reward_model_maze2d-medium-v0'}, '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': 1, 'n_action_steps': 1, '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-general-freward.pt', 'data_path2': 'datasets/maze2d-medium-v0-freward.pt', 'horizon': 1, 'pad_before': 0, 'pad_after': 0, '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'}}}
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():669] starting backend
+2026-07-25 15:13:03,719 INFO    MainThread:707073 [wandb_init.py:init():673] sending inform_init request
+2026-07-25 15:13:03,721 INFO    MainThread:707073 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-07-25 15:13:03,722 INFO    MainThread:707073 [wandb_init.py:init():686] backend started and connected
+2026-07-25 15:13:03,729 INFO    MainThread:707073 [wandb_init.py:init():781] updated telemetry
+2026-07-25 15:13:03,762 INFO    MainThread:707073 [wandb_init.py:init():814] communicating run to backend with 90.0 second timeout
+2026-07-25 15:13:04,591 INFO    MainThread:707073 [wandb_init.py:init():867] starting run threads in backend
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_console_start():2451] atexit reg
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2299] redirect: wrap_raw
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2364] Wrapping output streams.
+2026-07-25 15:13:04,775 INFO    MainThread:707073 [wandb_run.py:_redirect():2389] Redirects installed.
+2026-07-25 15:13:04,777 INFO    MainThread:707073 [wandb_init.py:init():911] run started, returning control to user process
+2026-07-25 15:13:04,777 INFO    MainThread:707073 [wandb_run.py:_config_callback():1389] config_cb None None {'output_dir': '/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0'}
+2026-07-25 17:07:31,921 WARNING MsgRouterThr:707073 [router.py:message_loop():75] message_loop has been closed

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/run-20260725_151303-i12soqic/run-i12soqic.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0093c2da12207b6afd0c3a1a04b79713dbd252801771e3bb11c9262a00ec5b48
+size 620328106

2026.07.25/15.12.59_train_rbc_reward_model_maze2d-medium-v0/wandb/wandb-resume.json

@@ -0,0 +1 @@
+{"run_id": "i12soqic"}