Upload folder using huggingface_hub

.gitattributes

@@ -258,3 +258,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 2025.12.08/11.27.59_train_llmbc_lowdim_box-close-v2/wandb/run-20251208_112809-rtrw2cf6/run-rtrw2cf6.wandb filter=lfs diff=lfs merge=lfs -text
 2025.12.01/17.47.28_train_bc_lowdim_box-close-v2/wandb/run-20251201_174741-8ciqmy4q/run-8ciqmy4q.wandb filter=lfs diff=lfs merge=lfs -text
 2026.07.26/02.46.34_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024637-9hdne6gf/run-9hdne6gf.wandb filter=lfs diff=lfs merge=lfs -text
+2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/run-fpoudnpx.wandb filter=lfs diff=lfs merge=lfs -text

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/.hydra/config.yaml

@@ -0,0 +1,112 @@
+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:
+  - 128
+  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: 0
+  num_epochs: 200
+  gradient_accumulate_every: 1
+  grad_norm_clip: 0.5
+  checkpoint_every: 10
+  val_every: 1
+  sample_every: 10
+  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: adroit-hand-hammer-v1
+  obs_dim: 46
+  action_dim: 26
+  env_runner:
+    _target_: llmbc.env_runner.adroit_lowdim_runner.AdroitHandLowdimRunner
+    env_name: llf-adroit-adroit-hand-hammer-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-hammer-v1-general-freward.pt
+    data_path2: datasets/adroit-hand-hammer-v1-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: 0.1
+    dummy_normalizer: false
+  instructor:
+    _target_: llmbc.translator.instructor.adroit_instructor.adroit_hand_hammer_v1_instructor.AdroitHandHammerV1Instructor

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-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: rdp_reward-freward_adroit-hand-hammer-v1.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/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-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

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/.hydra/overrides.yaml

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

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/epoch=0130-val_loss=0.068.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ae957ead65f67dc3c716d3ec48da2ef212c576f1bf3d4e556db52f56d2f921e
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/epoch=0140-val_loss=0.069.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:444251cbdd42824956c98c938d083c8aac39b7f67731dce6ae61caf5a67b0c26
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/epoch=0160-val_loss=0.069.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1e1707c79a86a7806215d7bad4554141b514fdd9a9118dde0b3c6272fa8ffe7
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/epoch=0180-val_loss=0.068.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13c709afe00c0f0757c9c8cea975cd576a8a2bd273f70aa8f015f380a0522897
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/epoch=0190-val_loss=0.068.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34630a65efb6e7845871349d9a46417c6fa2e97da7632f0865ac44685146d1e0
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/checkpoints/latest.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34630a65efb6e7845871349d9a46417c6fa2e97da7632f0865ac44685146d1e0
+size 139388

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/train.log

@@ -0,0 +1,2 @@
+[2026-07-26 02:48:49,820][numexpr.utils][INFO] - Note: NumExpr detected 24 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
+[2026-07-26 02:48:49,820][numexpr.utils][INFO] - NumExpr defaulting to 8 threads.

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2026-07-26T02:48:52.526873342+08:00","level":"INFO","msg":"using version","core version":"0.18.6"}
+{"time":"2026-07-26T02:48:52.52688781+08:00","level":"INFO","msg":"created symlink","path":"/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-core.log"}
+{"time":"2026-07-26T02:48:52.632370294+08:00","level":"INFO","msg":"created new stream","id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.632402991+08:00","level":"INFO","msg":"stream: started","id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.633523173+08:00","level":"INFO","msg":"sender: started","stream_id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.633575463+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:48:52.633589872+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:48:53.806230067+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-07-26T02:50:43.338191876+08:00","level":"INFO","msg":"stream: closing","id":"fpoudnpx"}
+{"time":"2026-07-26T02:50:43.338222327+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-07-26T02:50:43.338785074+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-07-26T02:50:44.862642873+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-07-26T02:50:45.489229256+08:00","level":"INFO","msg":"handler: closed","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:50:45.489308001+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:50:45.489351341+08:00","level":"INFO","msg":"sender: closed","stream_id":"fpoudnpx"}
+{"time":"2026-07-26T02:50:45.489358773+08:00","level":"INFO","msg":"stream: closed","id":"fpoudnpx"}

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/debug.log

@@ -0,0 +1,27 @@
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Current SDK version is 0.18.6
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Configure stats pid to 2718622
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from /home/chyang/.config/wandb/settings
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from /tmp2/chyang/workspace/LLM-BC/wandb/settings
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': 'online', '_disable_service': None}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [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'}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:_log_setup():533] Logging user logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug.log
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:_log_setup():534] Logging internal logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-internal.log
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:init():619] calling init triggers
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [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': 46, 'action_dim': 26, 'task_name': 'adroit-hand-hammer-v1', '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': 72, 'hidden_size': [128], 'obs_dim': 46, 'action_dim': 26, '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': 0, 'num_epochs': 200, 'gradient_accumulate_every': 1, 'grad_norm_clip': 0.5, 'checkpoint_every': 10, 'val_every': 1, 'sample_every': 10, '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': '2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1', 'tags': ['train_rbc_reward_model', 'adroit-hand-hammer-v1', '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/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1', 'wandb_name_base': '2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1'}, 'task': {'name': 'adroit-hand-hammer-v1', 'obs_dim': 46, 'action_dim': 26, 'env_runner': {'_target_': 'llmbc.env_runner.adroit_lowdim_runner.AdroitHandLowdimRunner', 'env_name': 'llf-adroit-adroit-hand-hammer-v1', 'n_train': 10, 'n_test': 50, 'n_envs': 10, 'max_steps': 150, '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.adroit_lowdim_dataset.AdroitHandLowdimDataset', 'data_path': 'datasets/adroit-hand-hammer-v1-general-freward.pt', 'data_path2': 'datasets/adroit-hand-hammer-v1-freward.pt', 'horizon': 1, 'pad_before': 0, 'pad_after': 0, 'obs_eef_target': True, 'use_manual_normalizer': False, 'val_ratio': 0.1, 'dummy_normalizer': False}, 'instructor': {'_target_': 'llmbc.translator.instructor.adroit_instructor.adroit_hand_hammer_v1_instructor.AdroitHandHammerV1Instructor'}}}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:init():669] starting backend
+2026-07-26 02:48:52,524 INFO    MainThread:2718622 [wandb_init.py:init():673] sending inform_init request
+2026-07-26 02:48:52,527 INFO    MainThread:2718622 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-07-26 02:48:52,527 INFO    MainThread:2718622 [wandb_init.py:init():686] backend started and connected
+2026-07-26 02:48:52,532 INFO    MainThread:2718622 [wandb_init.py:init():781] updated telemetry
+2026-07-26 02:48:52,564 INFO    MainThread:2718622 [wandb_init.py:init():814] communicating run to backend with 90.0 second timeout
+2026-07-26 02:48:53,793 INFO    MainThread:2718622 [wandb_init.py:init():867] starting run threads in backend
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_console_start():2451] atexit reg
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2299] redirect: wrap_raw
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2364] Wrapping output streams.
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2389] Redirects installed.
+2026-07-26 02:48:54,010 INFO    MainThread:2718622 [wandb_init.py:init():911] run started, returning control to user process
+2026-07-26 02:48:54,010 INFO    MainThread:2718622 [wandb_run.py:_config_callback():1389] config_cb None None {'output_dir': '/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1'}
+2026-07-26 02:50:43,338 WARNING MsgRouterThr:2718622 [router.py:message_loop():75] message_loop has been closed

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/files/config.yaml

@@ -0,0 +1,168 @@
+_target_:
+    value: llmbc.workspace.train_rbc_reward_workspace.TrainRBCRewardWorkspace
+_wandb:
+    value:
+        cli_version: 0.18.6
+        m: []
+        python_version: 3.9.20
+        t:
+            "1":
+                - 1
+                - 11
+                - 49
+                - 50
+                - 55
+                - 83
+            "2":
+                - 1
+                - 11
+                - 49
+                - 50
+                - 55
+                - 83
+            "3":
+                - 13
+                - 15
+                - 16
+                - 23
+                - 55
+                - 61
+            "4": 3.9.20
+            "5": 0.18.6
+            "6": 4.47.1
+            "8":
+                - 5
+            "12": 0.18.6
+            "13": linux-x86_64
+action_dim:
+    value: 26
+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: 2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1
+        project: rbc_module_training
+        resume: true
+        tags:
+            - train_rbc_reward_model
+            - adroit-hand-hammer-v1
+            - default
+multi_run:
+    value:
+        run_dir: data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1
+        wandb_name_base: 2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1
+n_action_steps:
+    value: 1
+n_latency_steps:
+    value: 0
+n_obs_steps:
+    value: 1
+name:
+    value: train_rbc_reward_model
+obs_dim:
+    value: 46
+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/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1
+past_action_visible:
+    value: false
+reward_model:
+    value:
+        _target_: llmbc.policy.rbc_lowdim_policy.RBCLowdimRewardModel
+        action_dim: 26
+        hidden_size:
+            - 128
+        input_size: 72
+        n_action_steps: 1
+        n_obs_steps: 1
+        obs_dim: 46
+task:
+    value:
+        action_dim: 26
+        dataset:
+            _target_: llmbc.dataset.adroit_lowdim_dataset.AdroitHandLowdimDataset
+            data_path: datasets/adroit-hand-hammer-v1-general-freward.pt
+            data_path2: datasets/adroit-hand-hammer-v1-freward.pt
+            dummy_normalizer: false
+            horizon: 1
+            obs_eef_target: true
+            pad_after: 0
+            pad_before: 0
+            use_manual_normalizer: false
+            val_ratio: 0.1
+        env_runner:
+            _target_: llmbc.env_runner.adroit_lowdim_runner.AdroitHandLowdimRunner
+            discount: 0.99
+            env_name: llf-adroit-adroit-hand-hammer-v1
+            feedback_type:
+                - hp
+                - hn
+                - fp
+            instruction_type: b
+            max_steps: 150
+            n_action_steps: 1
+            n_envs: 10
+            n_obs_steps: 1
+            n_test: 50
+            n_train: 10
+            visual: false
+        instructor:
+            _target_: llmbc.translator.instructor.adroit_instructor.adroit_hand_hammer_v1_instructor.AdroitHandHammerV1Instructor
+        name: adroit-hand-hammer-v1
+        obs_dim: 46
+task_name:
+    value: adroit-hand-hammer-v1
+training:
+    value:
+        checkpoint_every: 10
+        debug: false
+        device: cuda
+        grad_norm_clip: 0.5
+        gradient_accumulate_every: 1
+        lr_scheduler: cosine
+        lr_warmup_steps: 0
+        max_train_steps: null
+        max_val_steps: null
+        num_epochs: 200
+        resume: true
+        sample_every: 10
+        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

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/files/output.log

@@ -0,0 +1 @@
+                                                                                                                                                                                 

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/files/wandb-metadata.json

@@ -0,0 +1,55 @@
+{
+  "os": "Linux-6.8.0-101-generic-x86_64-with-glibc2.35",
+  "python": "3.9.20",
+  "startedAt": "2026-07-25T18:48:52.528073Z",
+  "args": [
+    "--config-path",
+    "config/main_table",
+    "--config-name",
+    "rdp_reward-freward_adroit-hand-hammer-v1.yaml"
+  ],
+  "program": "/tmp2/chyang/workspace/LLM-BC/./train.py",
+  "codePath": "train.py",
+  "git": {
+    "remote": "https://github.com/CHYang25/LLM-BC.git",
+    "commit": "7115513fceb8b6fa71504a5a121f22e33cae5873"
+  },
+  "email": "chris920325@gmail.com",
+  "root": "/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1",
+  "host": "A6000-2",
+  "username": "chyang",
+  "executable": "/home/chyang/miniconda3/envs/llm-bc/bin/python3",
+  "codePathLocal": "train.py",
+  "cpu_count": 12,
+  "cpu_count_logical": 24,
+  "gpu": "NVIDIA RTX A6000",
+  "gpu_count": 2,
+  "disk": {
+    "/": {
+      "total": "1967317549056",
+      "used": "1049034534912"
+    }
+  },
+  "memory": {
+    "total": "134538502144"
+  },
+  "cpu": {
+    "count": 12,
+    "countLogical": 24
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA RTX A6000",
+      "memoryTotal": "51527024640",
+      "cudaCores": 10752,
+      "architecture": "Ampere"
+    },
+    {
+      "name": "NVIDIA RTX A6000",
+      "memoryTotal": "51527024640",
+      "cudaCores": 10752,
+      "architecture": "Ampere"
+    }
+  ],
+  "cudaVersion": "12.6"
+}

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/files/wandb-summary.json

@@ -0,0 +1 @@
+{"epoch":199,"train_batch_mse_error":0.4389014306269708,"_runtime":110.810142736,"lr":0,"val_loss":0.06856992840766907,"_wandb":{"runtime":110},"train_loss":0.06858004744918574,"_timestamp":1.7850054433363903e+09,"_step":16799,"grad_norm":0.24438254535198212,"global_step":16799}

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-core.log

@@ -0,0 +1,14 @@
+{"time":"2026-07-26T02:48:51.907955806+08:00","level":"INFO","msg":"started logging, with flags","port-filename":"/tmp/tmpnepm1f2v/port-2718622.txt","pid":2718622,"debug":false,"disable-analytics":false}
+{"time":"2026-07-26T02:48:51.907995809+08:00","level":"INFO","msg":"FeatureState","shutdownOnParentExitEnabled":false}
+{"time":"2026-07-26T02:48:51.912007191+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":2718622}
+{"time":"2026-07-26T02:48:51.911991983+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":39919,"Zone":""}}
+{"time":"2026-07-26T02:48:52.097925162+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:48:52.526713301+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"fpoudnpx","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:48:52.632407884+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"fpoudnpx","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:43.338116187+08:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:43.338192101+08:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-07-26T02:50:43.338256555+08:00","level":"INFO","msg":"connection: Close: initiating connection closure","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:43.338312678+08:00","level":"INFO","msg":"connection: Close: connection successfully closed","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:45.489424005+08:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:45.489446424+08:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"127.0.0.1:36706"}
+{"time":"2026-07-26T02:50:45.489462672+08:00","level":"INFO","msg":"server is closed"}

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2026-07-26T02:48:52.526873342+08:00","level":"INFO","msg":"using version","core version":"0.18.6"}
+{"time":"2026-07-26T02:48:52.52688781+08:00","level":"INFO","msg":"created symlink","path":"/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-core.log"}
+{"time":"2026-07-26T02:48:52.632370294+08:00","level":"INFO","msg":"created new stream","id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.632402991+08:00","level":"INFO","msg":"stream: started","id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.633523173+08:00","level":"INFO","msg":"sender: started","stream_id":"fpoudnpx"}
+{"time":"2026-07-26T02:48:52.633575463+08:00","level":"INFO","msg":"handler: started","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:48:52.633589872+08:00","level":"INFO","msg":"writer: Do: started","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:48:53.806230067+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2026-07-26T02:50:43.338191876+08:00","level":"INFO","msg":"stream: closing","id":"fpoudnpx"}
+{"time":"2026-07-26T02:50:43.338222327+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2026-07-26T02:50:43.338785074+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2026-07-26T02:50:44.862642873+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-07-26T02:50:45.489229256+08:00","level":"INFO","msg":"handler: closed","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:50:45.489308001+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":{"value":"fpoudnpx"}}
+{"time":"2026-07-26T02:50:45.489351341+08:00","level":"INFO","msg":"sender: closed","stream_id":"fpoudnpx"}
+{"time":"2026-07-26T02:50:45.489358773+08:00","level":"INFO","msg":"stream: closed","id":"fpoudnpx"}

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug.log

@@ -0,0 +1,27 @@
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Current SDK version is 0.18.6
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Configure stats pid to 2718622
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from /home/chyang/.config/wandb/settings
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from /tmp2/chyang/workspace/LLM-BC/wandb/settings
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Loading settings from environment variables: {}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Applying setup settings: {'mode': 'online', '_disable_service': None}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [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'}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_setup.py:_flush():79] Applying login settings: {}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:_log_setup():533] Logging user logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug.log
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:_log_setup():534] Logging internal logs to /tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/logs/debug-internal.log
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:init():619] calling init triggers
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [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': 46, 'action_dim': 26, 'task_name': 'adroit-hand-hammer-v1', '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': 72, 'hidden_size': [128], 'obs_dim': 46, 'action_dim': 26, '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': 0, 'num_epochs': 200, 'gradient_accumulate_every': 1, 'grad_norm_clip': 0.5, 'checkpoint_every': 10, 'val_every': 1, 'sample_every': 10, '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': '2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1', 'tags': ['train_rbc_reward_model', 'adroit-hand-hammer-v1', '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/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1', 'wandb_name_base': '2026.07.26-02.48.49_train_rbc_reward_model_adroit-hand-hammer-v1'}, 'task': {'name': 'adroit-hand-hammer-v1', 'obs_dim': 46, 'action_dim': 26, 'env_runner': {'_target_': 'llmbc.env_runner.adroit_lowdim_runner.AdroitHandLowdimRunner', 'env_name': 'llf-adroit-adroit-hand-hammer-v1', 'n_train': 10, 'n_test': 50, 'n_envs': 10, 'max_steps': 150, '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.adroit_lowdim_dataset.AdroitHandLowdimDataset', 'data_path': 'datasets/adroit-hand-hammer-v1-general-freward.pt', 'data_path2': 'datasets/adroit-hand-hammer-v1-freward.pt', 'horizon': 1, 'pad_before': 0, 'pad_after': 0, 'obs_eef_target': True, 'use_manual_normalizer': False, 'val_ratio': 0.1, 'dummy_normalizer': False}, 'instructor': {'_target_': 'llmbc.translator.instructor.adroit_instructor.adroit_hand_hammer_v1_instructor.AdroitHandHammerV1Instructor'}}}
+2026-07-26 02:48:52,523 INFO    MainThread:2718622 [wandb_init.py:init():669] starting backend
+2026-07-26 02:48:52,524 INFO    MainThread:2718622 [wandb_init.py:init():673] sending inform_init request
+2026-07-26 02:48:52,527 INFO    MainThread:2718622 [backend.py:_multiprocessing_setup():104] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2026-07-26 02:48:52,527 INFO    MainThread:2718622 [wandb_init.py:init():686] backend started and connected
+2026-07-26 02:48:52,532 INFO    MainThread:2718622 [wandb_init.py:init():781] updated telemetry
+2026-07-26 02:48:52,564 INFO    MainThread:2718622 [wandb_init.py:init():814] communicating run to backend with 90.0 second timeout
+2026-07-26 02:48:53,793 INFO    MainThread:2718622 [wandb_init.py:init():867] starting run threads in backend
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_console_start():2451] atexit reg
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2299] redirect: wrap_raw
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2364] Wrapping output streams.
+2026-07-26 02:48:54,008 INFO    MainThread:2718622 [wandb_run.py:_redirect():2389] Redirects installed.
+2026-07-26 02:48:54,010 INFO    MainThread:2718622 [wandb_init.py:init():911] run started, returning control to user process
+2026-07-26 02:48:54,010 INFO    MainThread:2718622 [wandb_run.py:_config_callback():1389] config_cb None None {'output_dir': '/tmp2/chyang/workspace/LLM-BC/data/outputs/2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1'}
+2026-07-26 02:50:43,338 WARNING MsgRouterThr:2718622 [router.py:message_loop():75] message_loop has been closed

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/run-20260726_024852-fpoudnpx/run-fpoudnpx.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b902a60ba13955a62f866388333549362982fc006898cbf938b2f90ff744732b
+size 8819533

2026.07.26/02.48.48_train_rbc_reward_model_adroit-hand-hammer-v1/wandb/wandb-resume.json

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