Add gr00t/run_3_real (86 files)

gr00t/run_3_real/checkpoint-10000/config.json

@@ -0,0 +1,70 @@
+{
+  "action_horizon": 50,
+  "add_pos_embed": true,
+  "apply_sincos_state_encoding": true,
+  "architectures": [
+    "Gr00tN1d6"
+  ],
+  "attn_dropout": 0.2,
+  "attn_implementation": null,
+  "backbone_embedding_dim": 2048,
+  "backbone_model_type": "eagle",
+  "backbone_trainable_params_fp32": true,
+  "collator_overwrite_image_inputs": false,
+  "color_jitter_params": {
+    "brightness": 0.1,
+    "contrast": 0.1,
+    "hue": 0.1,
+    "saturation": 0.1
+  },
+  "crop_fraction": 0.95,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 48,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "interleave_self_attention": true,
+    "norm_type": "ada_norm",
+    "num_attention_heads": 32,
+    "num_layers": 32,
+    "output_dim": 1024,
+    "positional_embeddings": null
+  },
+  "eagle_collator": true,
+  "formalize_language": true,
+  "gemma_collator": false,
+  "hidden_size": 1024,
+  "image_crop_size": null,
+  "image_target_size": null,
+  "input_embedding_dim": 1536,
+  "load_bf16": true,
+  "max_action_dim": 128,
+  "max_num_embodiments": 32,
+  "max_seq_len": 1024,
+  "max_state_dim": 128,
+  "model_dtype": "bfloat16",
+  "model_name": "nvidia/Eagle-Block2A-2B-v2",
+  "model_type": "Gr00tN1d6",
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_inference_timesteps": 4,
+  "num_timestep_buckets": 1000,
+  "random_rotation_angle": null,
+  "reproject_vision": false,
+  "select_layer": 16,
+  "shortest_image_edge": 256,
+  "state_dropout_prob": 0.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "tune_diffusion_model": true,
+  "tune_llm": true,
+  "tune_projector": true,
+  "tune_top_llm_layers": 4,
+  "tune_visual": true,
+  "tune_vlln": true,
+  "use_albumentations_transforms": true,
+  "use_alternate_vl_dit": true,
+  "use_flash_attention": true,
+  "use_relative_action": true,
+  "use_vlln": true
+}

gr00t/run_3_real/checkpoint-10000/global_step10000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ae19756c20bb5324d9be29da81795020d9104d38f43ca1644b11f555361f4e2
+size 1476395008

gr00t/run_3_real/checkpoint-10000/global_step10000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:605dd7d1bf130eeb6016f9c659e3946940408d3449fa02bc97c085c0d45277ae
+size 7247757312

gr00t/run_3_real/checkpoint-10000/global_step10000/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54d41010b5368c7e9b782e0700a66e20092ec6992e34783848aa64ac1f0b3ad0
+size 6573616357

gr00t/run_3_real/checkpoint-10000/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97ac05f2aafa14af49c9f2323b9a4476bea72777935b3defd98dd79fcce09b78
+size 4991091456

gr00t/run_3_real/checkpoint-10000/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0680f1a908d671ac0541404e8d3d9c951be2a225b9511048b889c2a45b9289a
+size 1582283096

gr00t/run_3_real/checkpoint-10000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad069f2136738ad1d5be32c92812831e4eab7f84f59a54047d3d72150c0d6ae7
+size 7697

gr00t/run_3_real/checkpoint-3000/config.json

@@ -0,0 +1,70 @@
+{
+  "action_horizon": 50,
+  "add_pos_embed": true,
+  "apply_sincos_state_encoding": true,
+  "architectures": [
+    "Gr00tN1d6"
+  ],
+  "attn_dropout": 0.2,
+  "attn_implementation": null,
+  "backbone_embedding_dim": 2048,
+  "backbone_model_type": "eagle",
+  "backbone_trainable_params_fp32": true,
+  "collator_overwrite_image_inputs": false,
+  "color_jitter_params": {
+    "brightness": 0.1,
+    "contrast": 0.1,
+    "hue": 0.1,
+    "saturation": 0.1
+  },
+  "crop_fraction": 0.95,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 48,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "interleave_self_attention": true,
+    "norm_type": "ada_norm",
+    "num_attention_heads": 32,
+    "num_layers": 32,
+    "output_dim": 1024,
+    "positional_embeddings": null
+  },
+  "eagle_collator": true,
+  "formalize_language": true,
+  "gemma_collator": false,
+  "hidden_size": 1024,
+  "image_crop_size": null,
+  "image_target_size": null,
+  "input_embedding_dim": 1536,
+  "load_bf16": true,
+  "max_action_dim": 128,
+  "max_num_embodiments": 32,
+  "max_seq_len": 1024,
+  "max_state_dim": 128,
+  "model_dtype": "bfloat16",
+  "model_name": "nvidia/Eagle-Block2A-2B-v2",
+  "model_type": "Gr00tN1d6",
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_inference_timesteps": 4,
+  "num_timestep_buckets": 1000,
+  "random_rotation_angle": null,
+  "reproject_vision": false,
+  "select_layer": 16,
+  "shortest_image_edge": 256,
+  "state_dropout_prob": 0.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "tune_diffusion_model": true,
+  "tune_llm": true,
+  "tune_projector": true,
+  "tune_top_llm_layers": 4,
+  "tune_visual": true,
+  "tune_vlln": true,
+  "use_albumentations_transforms": true,
+  "use_alternate_vl_dit": true,
+  "use_flash_attention": true,
+  "use_relative_action": true,
+  "use_vlln": true
+}

gr00t/run_3_real/checkpoint-3000/embodiment_id.json

@@ -0,0 +1,11 @@
+{
+  "robocasa_panda_omron": 13,
+  "gr1": 20,
+  "behavior_r1_pro": 24,
+  "unitree_g1": 8,
+  "oxe_google": 0,
+  "oxe_widowx": 1,
+  "libero_panda": 2,
+  "oxe_droid": 16,
+  "new_embodiment": 10
+}

gr00t/run_3_real/checkpoint-3000/experiment_cfg/conf.yaml

@@ -0,0 +1,246 @@
+load_config_path: null
+model:
+  model_type: Gr00tN1d6
+  model_dtype: bfloat16
+  model_name: nvidia/Eagle-Block2A-2B-v2
+  backbone_model_type: eagle
+  model_revision: null
+  tune_top_llm_layers: 4
+  backbone_embedding_dim: 2048
+  tune_llm: true
+  tune_visual: true
+  select_layer: 16
+  reproject_vision: false
+  use_flash_attention: true
+  load_bf16: false
+  collator_overwrite_image_inputs: false
+  eagle_collator: true
+  backbone_trainable_params_fp32: true
+  image_crop_size: null
+  image_target_size: null
+  shortest_image_edge: 256
+  crop_fraction: 0.95
+  random_rotation_angle: null
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    saturation: 0.5
+    hue: 0.08
+  use_albumentations_transforms: true
+  formalize_language: true
+  apply_sincos_state_encoding: false
+  use_relative_action: true
+  max_state_dim: 29
+  max_action_dim: 29
+  action_horizon: 16
+  hidden_size: 1024
+  input_embedding_dim: 1536
+  add_pos_embed: true
+  attn_dropout: 0.2
+  use_vlln: true
+  max_seq_len: 1024
+  use_alternate_vl_dit: true
+  attend_text_every_n_blocks: 2
+  diffusion_model_cfg:
+    positional_embeddings: null
+    num_layers: 32
+    num_attention_heads: 32
+    attention_head_dim: 48
+    norm_type: ada_norm
+    dropout: 0.2
+    final_dropout: true
+    output_dim: 1024
+    interleave_self_attention: true
+  num_inference_timesteps: 4
+  noise_beta_alpha: 1.5
+  noise_beta_beta: 1.0
+  noise_s: 0.999
+  num_timestep_buckets: 1000
+  tune_projector: true
+  tune_diffusion_model: true
+  tune_vlln: true
+  state_dropout_prob: 0.0
+  state_additive_noise_scale: 0.0
+  max_num_embodiments: 32
+data:
+  datasets:
+  - dataset_paths:
+    - /workspace/capstone-vla/dataset_tools/data/realWorldPouring
+    embodiment_tag: new_embodiment
+    mix_ratio: 1.0
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  modality_configs:
+    new_embodiment:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - ego_view
+        - left_wrist_view
+        - right_wrist_view
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+  download_cache: false
+  shard_size: 1024
+  episode_sampling_rate: 0.1
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  mode: single_turn
+  random_chop: 0.0
+  mock_dataset_mode: false
+  shuffle: true
+  seed: 42
+  multiprocessing_context: fork
+  allow_padding: false
+  subsample_ratio: 1.0
+  image_crop_size:
+  - 244
+  - 244
+  image_target_size:
+  - 224
+  - 224
+  video_backend: torchcodec
+training:
+  output_dir: /workspace/capstone-vla/third_party/gr00t/chkpt_rlwd/run_3_real
+  experiment_name: null
+  max_steps: 10000
+  global_batch_size: 32
+  batch_size: null
+  gradient_accumulation_steps: 1
+  learning_rate: 0.0001
+  lr_scheduler_type: cosine
+  weight_decay: 1.0e-05
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  max_grad_norm: 1.0
+  optim: adamw_torch
+  start_from_checkpoint: nvidia/GR00T-N1.6-3B
+  tf32: true
+  fp16: false
+  bf16: true
+  eval_bf16: true
+  logging_steps: 10
+  save_steps: 3000
+  save_total_limit: 3
+  save_vl_model: false
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  max_concurrent_uploads: 2
+  eval_strategy: 'no'
+  eval_steps: 500
+  eval_set_split_ratio: 0.1
+  eval_batch_size: 2
+  save_best_eval_metric_name: ''
+  save_best_eval_metric_greater_is_better: true
+  deepspeed_stage: 2
+  gradient_checkpointing: false
+  transformers_trust_remote_code: true
+  transformers_local_files_only: false
+  transformers_cache_dir: null
+  transformers_access_token: null
+  use_ddp: false
+  ddp_bucket_cap_mb: 100
+  num_gpus: 2
+  dataloader_num_workers: 4
+  remove_unused_columns: false
+  use_wandb: true
+  wandb_project: finetune-gr00t-n1d6
+  enable_profiling: false
+  max_retries: 3
+  assert_loss_less_than: null
+  add_rl_callback: false
+  enable_open_loop_eval: false
+  open_loop_eval_traj_ids:
+  - 0
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_plot_indices: null
+max_steps: 10000
+save_steps: 3000

gr00t/run_3_real/checkpoint-3000/experiment_cfg/config.yaml

@@ -0,0 +1,277 @@
+!!python/object:gr00t.configs.base_config.Config
+data: !!python/object:gr00t.configs.data.data_config.DataConfig
+  allow_padding: false
+  datasets:
+  - !!python/object:gr00t.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /workspace/capstone-vla/dataset_tools/data/realWorldPouring
+    dataset_type: physical_embodiment
+    embodiment_tag: new_embodiment
+    mix_ratio: 1.0
+    val_dataset_path: null
+  download_cache: false
+  episode_sampling_rate: 0.1
+  image_crop_size:
+  - 244
+  - 244
+  image_target_size:
+  - 224
+  - 224
+  mock_dataset_mode: false
+  modality_configs:
+    new_embodiment:
+      action: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: &id001 !!python/object/apply:gr00t.data.types.ActionFormat
+          - default
+          rep: &id002 !!python/object/apply:gr00t.data.types.ActionRepresentation
+          - absolute
+          state_key: null
+          type: &id003 !!python/object/apply:gr00t.data.types.ActionType
+          - non_eef
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys: &id004
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+      language: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys: *id004
+        sin_cos_embedding_keys: null
+      video: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - ego_view
+        - left_wrist_view
+        - right_wrist_view
+        sin_cos_embedding_keys: null
+  mode: single_turn
+  multiprocessing_context: fork
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  random_chop: 0.0
+  seed: 42
+  shard_size: 1024
+  shuffle: true
+  subsample_ratio: 1.0
+  video_backend: torchcodec
+load_config_path: null
+model: !!python/object:gr00t.configs.model.gr00t_n1d6.Gr00tN1d6Config
+  _attn_implementation_autoset: false
+  _attn_implementation_internal: null
+  _commit_hash: null
+  _name_or_path: ''
+  add_cross_attention: false
+  architectures: null
+  backbone_model_type: eagle
+  backbone_trainable_params_fp32: true
+  bad_words_ids: null
+  begin_suppress_tokens: null
+  bos_token_id: null
+  chunk_size_feed_forward: 0
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    hue: 0.08
+    saturation: 0.5
+  cross_attention_hidden_size: null
+  decoder_start_token_id: null
+  diffusion_model_cfg:
+    attention_head_dim: 48
+    dropout: 0.2
+    final_dropout: true
+    interleave_self_attention: true
+    norm_type: ada_norm
+    num_attention_heads: 32
+    num_layers: 32
+    output_dim: 1024
+    positional_embeddings: null
+  diversity_penalty: 0.0
+  do_sample: false
+  eagle_collator: true
+  early_stopping: false
+  encoder_no_repeat_ngram_size: 0
+  eos_token_id: null
+  exponential_decay_length_penalty: null
+  finetuning_task: null
+  forced_bos_token_id: null
+  forced_eos_token_id: null
+  id2label:
+    0: LABEL_0
+    1: LABEL_1
+  is_decoder: false
+  is_encoder_decoder: false
+  label2id:
+    LABEL_0: 0
+    LABEL_1: 1
+  length_penalty: 1.0
+  load_bf16: false
+  max_length: 20
+  min_length: 0
+  model_name: nvidia/Eagle-Block2A-2B-v2
+  no_repeat_ngram_size: 0
+  num_beam_groups: 1
+  num_beams: 1
+  num_return_sequences: 1
+  output_attentions: false
+  output_hidden_states: false
+  output_scores: false
+  pad_token_id: null
+  prefix: null
+  problem_type: null
+  pruned_heads: {}
+  random_rotation_angle: null
+  remove_invalid_values: false
+  repetition_penalty: 1.0
+  reproject_vision: false
+  return_dict: true
+  return_dict_in_generate: false
+  sep_token_id: null
+  state_dropout_prob: 0.0
+  suppress_tokens: null
+  task_specific_params: null
+  temperature: 1.0
+  tf_legacy_loss: false
+  tie_encoder_decoder: false
+  tie_word_embeddings: true
+  tokenizer_class: null
+  top_k: 50
+  top_p: 1.0
+  torch_dtype: null
+  torchscript: false
+  transformers_version: null
+  tune_diffusion_model: true
+  tune_llm: true
+  tune_projector: true
+  tune_visual: true
+  typical_p: 1.0
+  use_bfloat16: false
+  use_relative_action: true
+training: !!python/object:gr00t.configs.training.training_config.TrainingConfig
+  add_rl_callback: false
+  assert_loss_less_than: null
+  batch_size: null
+  bf16: true
+  dataloader_num_workers: 4
+  ddp_bucket_cap_mb: 100
+  deepspeed_stage: 2
+  enable_open_loop_eval: false
+  enable_profiling: false
+  eval_batch_size: 2
+  eval_bf16: true
+  eval_set_split_ratio: 0.1
+  eval_steps: 500
+  eval_strategy: 'no'
+  experiment_name: null
+  fp16: false
+  global_batch_size: 32
+  gradient_accumulation_steps: 1
+  gradient_checkpointing: false
+  learning_rate: 0.0001
+  logging_steps: 10
+  lr_scheduler_type: cosine
+  max_concurrent_uploads: 2
+  max_grad_norm: 1.0
+  max_retries: 3
+  max_steps: 10000
+  num_gpus: 2
+  open_loop_eval_plot_indices: null
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_traj_ids:
+  - 0
+  optim: adamw_torch
+  output_dir: /workspace/capstone-vla/third_party/gr00t/chkpt_rlwd/run_3_real
+  remove_unused_columns: false
+  save_best_eval_metric_greater_is_better: true
+  save_best_eval_metric_name: ''
+  save_steps: 3000
+  save_total_limit: 3
+  save_vl_model: false
+  start_from_checkpoint: nvidia/GR00T-N1.6-3B
+  tf32: true
+  transformers_access_token: null
+  transformers_cache_dir: null
+  transformers_local_files_only: false
+  transformers_trust_remote_code: true
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  use_ddp: false
+  use_wandb: true
+  wandb_project: finetune-gr00t-n1d6
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  weight_decay: 1.0e-05

gr00t/run_3_real/checkpoint-3000/experiment_cfg/dataset_statistics.json

@@ -0,0 +1,761 @@
+{
+  "new_embodiment": {
+    "state": {
+      "left_arm": {
+        "min": [
+          -1.4764578342437744,
+          -0.1518283635377884,
+          -1.499962568283081,
+          -2.27626633644104,
+          -1.5671002864837646,
+          -0.6656755208969116,
+          -0.5207215547561646
+        ],
+        "max": [
+          0.3441833555698395,
+          0.657663881778717,
+          0.04959612339735031,
+          -0.7633329629898071,
+          0.5798436999320984,
+          0.6755949258804321,
+          0.6866571307182312
+        ],
+        "mean": [
+          -0.24634727835655212,
+          0.2629535496234894,
+          -0.5132859945297241,
+          -1.822871208190918,
+          -0.15705835819244385,
+          0.250943660736084,
+          -0.009801508858799934
+        ],
+        "std": [
+          0.2749035060405731,
+          0.061663031578063965,
+          0.23117874562740326,
+          0.2519250214099884,
+          0.42646297812461853,
+          0.20134037733078003,
+          0.23961548507213593
+        ],
+        "q01": [
+          -1.1976326751708983,
+          0.1270393806695938,
+          -1.415273575782776,
+          -2.1694675636291505,
+          -1.555278000831604,
+          -0.4085486590862274,
+          -0.3467614698410034
+        ],
+        "q99": [
+          0.11830186098813932,
+          0.4688370370864868,
+          -0.08621232867240906,
+          -1.1188685512542724,
+          0.2598086118698119,
+          0.6740685701370239,
+          0.5260765767097467
+        ]
+      },
+      "left_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.7819951176643372,
+          -0.15371035039424896,
+          -0.7819951176643372,
+          -0.7819951176643372,
+          -1.5639902353286743,
+          -1.5639902353286743
+        ],
+        "std": [
+          1.2748703956604004,
+          1.4928919076919556,
+          1.2748703956604004,
+          1.2748703956604004,
+          2.549740791320801,
+          2.549740791320801
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "left_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "neck": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "right_arm": {
+        "min": [
+          -0.7119174599647522,
+          -0.55275958776474,
+          -0.09536845237016678,
+          -2.144679546356201,
+          -0.6652951240539551,
+          -0.5672546625137329,
+          -0.5874713063240051
+        ],
+        "max": [
+          0.33939826488494873,
+          -0.002669741865247488,
+          0.928513765335083,
+          -0.8445846438407898,
+          1.433588981628418,
+          0.4833299517631531,
+          0.6599544286727905
+        ],
+        "mean": [
+          -0.21578556299209595,
+          -0.22310534119606018,
+          0.40323933959007263,
+          -1.6887274980545044,
+          0.08409612625837326,
+          -0.18936818838119507,
+          -0.007405608426779509
+        ],
+        "std": [
+          0.16777803003787994,
+          0.07015040516853333,
+          0.16941194236278534,
+          0.19395458698272705,
+          0.5333251953125,
+          0.1218024268746376,
+          0.20256562530994415
+        ],
+        "q01": [
+          -0.5456185221672059,
+          -0.4390794897079468,
+          0.054558104127645494,
+          -2.009228744506836,
+          -0.5162186360359191,
+          -0.4283971929550171,
+          -0.3479098403453827
+        ],
+        "q99": [
+          0.16945502102374999,
+          -0.08812180131673814,
+          0.7629516124725342,
+          -1.1490095472335817,
+          1.3221839523315426,
+          0.1579314386844635,
+          0.45433507800102235
+        ]
+      },
+      "right_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.41444578766822815,
+          0.15046590566635132,
+          -0.41444578766822815,
+          -0.41444578766822815,
+          -0.8288915753364563,
+          -0.8288915753364563
+        ],
+        "std": [
+          1.432570219039917,
+          1.4905503988265991,
+          1.432570219039917,
+          1.432570219039917,
+          2.865140438079834,
+          2.865140438079834
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "right_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "waist": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      }
+    },
+    "action": {
+      "left_arm": {
+        "min": [
+          -1.5419785976409912,
+          -0.19808706641197205,
+          -1.5707963705062866,
+          -2.3319737911224365,
+          -1.5707963705062866,
+          -0.6781747937202454,
+          -0.5912933349609375
+        ],
+        "max": [
+          0.339247465133667,
+          0.6958394646644592,
+          0.09750956296920776,
+          -0.7315306067466736,
+          0.6010454297065735,
+          0.6981316804885864,
+          0.6981316804885864
+        ],
+        "mean": [
+          -0.27577918767929077,
+          0.265458881855011,
+          -0.5064833760261536,
+          -1.8468353748321533,
+          -0.1631959080696106,
+          0.25038862228393555,
+          -0.011975681409239769
+        ],
+        "std": [
+          0.284289687871933,
+          0.06655233353376389,
+          0.23614443838596344,
+          0.26587340235710144,
+          0.432668536901474,
+          0.20466120541095734,
+          0.24292133748531342
+        ],
+        "q01": [
+          -1.254055471420288,
+          0.11745307803153991,
+          -1.4238929557800293,
+          -2.2081959438323975,
+          -1.5707963705062866,
+          -0.41540132522583006,
+          -0.3490602993965149
+        ],
+        "q99": [
+          0.09790542602539062,
+          0.48043997287750184,
+          -0.06798712402582203,
+          -1.1039583587646489,
+          0.2601294279098509,
+          0.6981316804885864,
+          0.53674639225006
+        ]
+      },
+      "left_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.7821466326713562,
+          -0.15372000634670258,
+          -0.7821466326713562,
+          -0.7821466326713562,
+          -1.5642932653427124,
+          -1.5642932653427124
+        ],
+        "std": [
+          1.2747758626937866,
+          1.4928909540176392,
+          1.2747758626937866,
+          1.2747758626937866,
+          2.5495517253875732,
+          2.5495517253875732
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "left_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "neck": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "right_arm": {
+        "min": [
+          -0.7609559297561646,
+          -0.5897848010063171,
+          -0.13661164045333862,
+          -2.1815736293792725,
+          -0.6809128522872925,
+          -0.5955591201782227,
+          -0.6164834499359131
+        ],
+        "max": [
+          0.31390857696533203,
+          0.01057685911655426,
+          0.9158902764320374,
+          -0.8264364004135132,
+          1.5061734914779663,
+          0.49332451820373535,
+          0.6635355353355408
+        ],
+        "mean": [
+          -0.24524866044521332,
+          -0.2248098999261856,
+          0.39028772711753845,
+          -1.7206827402114868,
+          0.08302607387304306,
+          -0.18993926048278809,
+          -0.012360971421003342
+        ],
+        "std": [
+          0.17419038712978363,
+          0.07439576834440231,
+          0.17066127061843872,
+          0.2032645046710968,
+          0.5400993824005127,
+          0.124636709690094,
+          0.20725084841251373
+        ],
+        "q01": [
+          -0.5854583382606506,
+          -0.45275834798812864,
+          0.03628820136189461,
+          -2.0545105361938476,
+          -0.5252219843864441,
+          -0.44349650502204896,
+          -0.36548688769340515
+        ],
+        "q99": [
+          0.15696598112583027,
+          -0.07897948294878006,
+          0.7490021729469298,
+          -1.1520879220962532,
+          1.3330879974365206,
+          0.15911160767078353,
+          0.4541356778144836
+        ]
+      },
+      "right_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.414561003446579,
+          0.15042728185653687,
+          -0.414561003446579,
+          -0.414561003446579,
+          -0.829122006893158,
+          -0.829122006893158
+        ],
+        "std": [
+          1.4325326681137085,
+          1.4905540943145752,
+          1.4325326681137085,
+          1.4325326681137085,
+          2.865065336227417,
+          2.865065336227417
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "right_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "waist": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      }
+    },
+    "relative_action": {}
+  }
+}

gr00t/run_3_real/checkpoint-3000/experiment_cfg/final_model_config.json

@@ -0,0 +1,53 @@
+{
+  "model_type": "Gr00tN1d6",
+  "model_dtype": "bfloat16",
+  "model_name": "nvidia/Eagle-Block2A-2B-v2",
+  "backbone_model_type": "eagle",
+  "model_revision": null,
+  "tune_top_llm_layers": 4,
+  "backbone_embedding_dim": 2048,
+  "tune_llm": true,
+  "tune_visual": true,
+  "select_layer": 16,
+  "reproject_vision": false,
+  "use_flash_attention": true,
+  "load_bf16": true,
+  "collator_overwrite_image_inputs": false,
+  "eagle_collator": true,
+  "backbone_trainable_params_fp32": true,
+  "apply_sincos_state_encoding": true,
+  "use_relative_action": true,
+  "max_state_dim": 128,
+  "max_action_dim": 128,
+  "action_horizon": 50,
+  "hidden_size": 1024,
+  "input_embedding_dim": 1536,
+  "add_pos_embed": true,
+  "attn_dropout": 0.2,
+  "use_vlln": true,
+  "max_seq_len": 1024,
+  "use_alternate_vl_dit": true,
+  "attend_text_every_n_blocks": 2,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 48,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "interleave_self_attention": true,
+    "norm_type": "ada_norm",
+    "num_attention_heads": 32,
+    "num_layers": 32,
+    "output_dim": 1024,
+    "positional_embeddings": null
+  },
+  "num_inference_timesteps": 4,
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_timestep_buckets": 1000,
+  "tune_projector": true,
+  "tune_diffusion_model": true,
+  "tune_vlln": true,
+  "state_dropout_prob": 0.0,
+  "state_additive_noise_scale": 0.0,
+  "max_num_embodiments": 32
+}

gr00t/run_3_real/checkpoint-3000/global_step3000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2a70dc86bf1cf38f0d79f25673262230ba29c15505bb838f8eb61299ed8e00c
+size 19719732513

gr00t/run_3_real/checkpoint-3000/global_step3000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3234c6cd10cacaad06e50a39f62bf895a37e9f75dfc6e6caff112f2f9b486bdf
+size 19719703265

gr00t/run_3_real/checkpoint-3000/global_step3000/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca7c5970f3038060c377016932a0c262f1ced2bd77cd475ca9c069e15bbb8801
+size 6573616357

gr00t/run_3_real/checkpoint-3000/latest

@@ -0,0 +1 @@
+global_step3000

gr00t/run_3_real/checkpoint-3000/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72c626e60cfaf869d92f83f3c159601dca784f5586f0ca181525fa81861ed71e
+size 4991091456

gr00t/run_3_real/checkpoint-3000/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528d898fe0f0d8bb31b535c4273f796dcfc0e23be05703340854ea01fcd5f17f
+size 1582283096

gr00t/run_3_real/checkpoint-3000/processor_config.json

@@ -0,0 +1,504 @@
+{
+  "processor_class": "Gr00tN1d6Processor",
+  "processor_kwargs": {
+    "modality_configs": {
+      "behavior_r1_pro": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "observation.images.rgb.head_256_256",
+            "observation.images.rgb.left_wrist_256_256",
+            "observation.images.rgb.right_wrist_256_256"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "robot_pos",
+            "robot_ori_cos",
+            "robot_ori_sin",
+            "robot_2d_ori",
+            "robot_2d_ori_cos",
+            "robot_2d_ori_sin",
+            "robot_lin_vel",
+            "robot_ang_vel",
+            "arm_left_qpos",
+            "arm_left_qpos_sin",
+            "arm_left_qpos_cos",
+            "eef_left_pos",
+            "eef_left_quat",
+            "gripper_left_qpos",
+            "arm_right_qpos",
+            "arm_right_qpos_sin",
+            "arm_right_qpos_cos",
+            "eef_right_pos",
+            "eef_right_quat",
+            "gripper_right_qpos",
+            "trunk_qpos"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15,
+            16,
+            17,
+            18,
+            19,
+            20,
+            21,
+            22,
+            23,
+            24,
+            25,
+            26,
+            27,
+            28,
+            29,
+            30,
+            31
+          ],
+          "modality_keys": [
+            "base",
+            "torso",
+            "left_arm",
+            "left_gripper",
+            "right_arm",
+            "right_gripper"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "trunk_qpos"
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "arm_left_qpos"
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "arm_right_qpos"
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.coarse_action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "gr1": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "ego_view_bg_crop_pad_res256_freq20"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "task"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "robocasa_panda_omron": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "res256_image_side_0",
+            "res256_image_side_1",
+            "res256_image_wrist_0"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position_relative",
+            "end_effector_rotation_relative",
+            "gripper_qpos",
+            "base_position",
+            "base_rotation"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close",
+            "base_motion",
+            "control_mode"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "new_embodiment": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "ego_view",
+            "left_wrist_view",
+            "right_wrist_view"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      }
+    },
+    "image_crop_size": null,
+    "image_target_size": null,
+    "use_albumentations": true,
+    "random_rotation_angle": null,
+    "color_jitter_params": {
+      "brightness": 0.3,
+      "contrast": 0.4,
+      "saturation": 0.5,
+      "hue": 0.08
+    },
+    "shortest_image_edge": 256,
+    "crop_fraction": 0.95,
+    "model_name": "nvidia/Eagle-Block2A-2B-v2",
+    "model_type": "eagle",
+    "formalize_language": true,
+    "max_state_dim": 128,
+    "max_action_dim": 128,
+    "max_action_horizon": 50,
+    "use_percentiles": false,
+    "clip_outliers": true,
+    "apply_sincos_state_encoding": true,
+    "use_relative_action": true
+  }
+}

gr00t/run_3_real/checkpoint-3000/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9854763a2d7a7d92546f45062c39be4b4234d839c4cad02886ed1cd91a6cdb88
+size 14917

gr00t/run_3_real/checkpoint-3000/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4610f47efa3ad7d951be51ae676f08f6e47b5ba3de478e65bc33afc49622e490
+size 14917

gr00t/run_3_real/checkpoint-3000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d434d3c9a7c15d35f1cd8e7e4a4476d02c75771edbbb965c02e4895df6eb318
+size 1465

gr00t/run_3_real/checkpoint-3000/trainer_state.json

@@ -0,0 +1,1834 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3,
+  "eval_steps": 500,
+  "global_step": 3000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "grad_norm": 1.450473666191101,
+      "learning_rate": 1.8e-06,
+      "loss": 1.3738,
+      "step": 10
+    },
+    {
+      "grad_norm": 0.9111957550048828,
+      "learning_rate": 3.8e-06,
+      "loss": 1.3531,
+      "step": 20
+    },
+    {
+      "grad_norm": 0.772489607334137,
+      "learning_rate": 5.8e-06,
+      "loss": 1.3273,
+      "step": 30
+    },
+    {
+      "grad_norm": 0.9024366140365601,
+      "learning_rate": 7.8e-06,
+      "loss": 1.291,
+      "step": 40
+    },
+    {
+      "grad_norm": 1.0620139837265015,
+      "learning_rate": 9.800000000000001e-06,
+      "loss": 1.2387,
+      "step": 50
+    },
+    {
+      "grad_norm": 1.9065420627593994,
+      "learning_rate": 1.18e-05,
+      "loss": 1.1773,
+      "step": 60
+    },
+    {
+      "grad_norm": 2.337477445602417,
+      "learning_rate": 1.3800000000000002e-05,
+      "loss": 1.1066,
+      "step": 70
+    },
+    {
+      "grad_norm": 5.197817802429199,
+      "learning_rate": 1.58e-05,
+      "loss": 1.0895,
+      "step": 80
+    },
+    {
+      "grad_norm": 2.817615270614624,
+      "learning_rate": 1.78e-05,
+      "loss": 1.0656,
+      "step": 90
+    },
+    {
+      "grad_norm": 1.9686524868011475,
+      "learning_rate": 1.9800000000000004e-05,
+      "loss": 1.0609,
+      "step": 100
+    },
+    {
+      "grad_norm": 0.8546046018600464,
+      "learning_rate": 2.18e-05,
+      "loss": 1.059,
+      "step": 110
+    },
+    {
+      "grad_norm": 1.7115342617034912,
+      "learning_rate": 2.38e-05,
+      "loss": 1.0613,
+      "step": 120
+    },
+    {
+      "grad_norm": 1.5974440574645996,
+      "learning_rate": 2.58e-05,
+      "loss": 1.0598,
+      "step": 130
+    },
+    {
+      "grad_norm": 0.7264371514320374,
+      "learning_rate": 2.7800000000000005e-05,
+      "loss": 1.0523,
+      "step": 140
+    },
+    {
+      "grad_norm": 1.0814321041107178,
+      "learning_rate": 2.98e-05,
+      "loss": 1.0477,
+      "step": 150
+    },
+    {
+      "grad_norm": 0.5987752676010132,
+      "learning_rate": 3.18e-05,
+      "loss": 1.0391,
+      "step": 160
+    },
+    {
+      "grad_norm": 0.49032509326934814,
+      "learning_rate": 3.38e-05,
+      "loss": 1.0361,
+      "step": 170
+    },
+    {
+      "grad_norm": 0.4149814546108246,
+      "learning_rate": 3.58e-05,
+      "loss": 1.0285,
+      "step": 180
+    },
+    {
+      "grad_norm": 0.46443986892700195,
+      "learning_rate": 3.7800000000000004e-05,
+      "loss": 1.0371,
+      "step": 190
+    },
+    {
+      "grad_norm": 0.43555909395217896,
+      "learning_rate": 3.9800000000000005e-05,
+      "loss": 1.0281,
+      "step": 200
+    },
+    {
+      "grad_norm": 0.45831650495529175,
+      "learning_rate": 4.18e-05,
+      "loss": 1.0344,
+      "step": 210
+    },
+    {
+      "grad_norm": 0.424280047416687,
+      "learning_rate": 4.38e-05,
+      "loss": 1.0227,
+      "step": 220
+    },
+    {
+      "grad_norm": 0.5616075992584229,
+      "learning_rate": 4.58e-05,
+      "loss": 1.0236,
+      "step": 230
+    },
+    {
+      "grad_norm": 0.4985102415084839,
+      "learning_rate": 4.78e-05,
+      "loss": 1.0064,
+      "step": 240
+    },
+    {
+      "grad_norm": 0.5166693329811096,
+      "learning_rate": 4.9800000000000004e-05,
+      "loss": 0.9885,
+      "step": 250
+    },
+    {
+      "grad_norm": 0.8828110694885254,
+      "learning_rate": 5.1800000000000005e-05,
+      "loss": 0.9826,
+      "step": 260
+    },
+    {
+      "grad_norm": 0.6461001038551331,
+      "learning_rate": 5.380000000000001e-05,
+      "loss": 0.9744,
+      "step": 270
+    },
+    {
+      "grad_norm": 0.7855983376502991,
+      "learning_rate": 5.580000000000001e-05,
+      "loss": 0.9553,
+      "step": 280
+    },
+    {
+      "grad_norm": 0.7762882709503174,
+      "learning_rate": 5.7799999999999995e-05,
+      "loss": 0.9254,
+      "step": 290
+    },
+    {
+      "grad_norm": 0.6695928573608398,
+      "learning_rate": 5.9800000000000003e-05,
+      "loss": 0.8973,
+      "step": 300
+    },
+    {
+      "grad_norm": 0.7439455389976501,
+      "learning_rate": 6.18e-05,
+      "loss": 0.8576,
+      "step": 310
+    },
+    {
+      "grad_norm": 0.9393029808998108,
+      "learning_rate": 6.38e-05,
+      "loss": 0.8107,
+      "step": 320
+    },
+    {
+      "grad_norm": 0.7973158359527588,
+      "learning_rate": 6.58e-05,
+      "loss": 0.7773,
+      "step": 330
+    },
+    {
+      "grad_norm": 0.8033375144004822,
+      "learning_rate": 6.780000000000001e-05,
+      "loss": 0.735,
+      "step": 340
+    },
+    {
+      "grad_norm": 0.9920246005058289,
+      "learning_rate": 6.98e-05,
+      "loss": 0.7123,
+      "step": 350
+    },
+    {
+      "grad_norm": 0.9083616733551025,
+      "learning_rate": 7.18e-05,
+      "loss": 0.6771,
+      "step": 360
+    },
+    {
+      "grad_norm": 0.8349876403808594,
+      "learning_rate": 7.38e-05,
+      "loss": 0.6408,
+      "step": 370
+    },
+    {
+      "grad_norm": 0.8299239873886108,
+      "learning_rate": 7.58e-05,
+      "loss": 0.6316,
+      "step": 380
+    },
+    {
+      "grad_norm": 0.9769677519798279,
+      "learning_rate": 7.780000000000001e-05,
+      "loss": 0.6186,
+      "step": 390
+    },
+    {
+      "grad_norm": 0.6790041923522949,
+      "learning_rate": 7.98e-05,
+      "loss": 0.5859,
+      "step": 400
+    },
+    {
+      "grad_norm": 0.7563514709472656,
+      "learning_rate": 8.18e-05,
+      "loss": 0.5746,
+      "step": 410
+    },
+    {
+      "grad_norm": 0.6624003052711487,
+      "learning_rate": 8.38e-05,
+      "loss": 0.5672,
+      "step": 420
+    },
+    {
+      "grad_norm": 0.8914895057678223,
+      "learning_rate": 8.58e-05,
+      "loss": 0.5481,
+      "step": 430
+    },
+    {
+      "grad_norm": 0.7826713919639587,
+      "learning_rate": 8.78e-05,
+      "loss": 0.5352,
+      "step": 440
+    },
+    {
+      "grad_norm": 0.72369784116745,
+      "learning_rate": 8.98e-05,
+      "loss": 0.5322,
+      "step": 450
+    },
+    {
+      "grad_norm": 0.6972424387931824,
+      "learning_rate": 9.180000000000001e-05,
+      "loss": 0.5126,
+      "step": 460
+    },
+    {
+      "grad_norm": 0.8057860732078552,
+      "learning_rate": 9.38e-05,
+      "loss": 0.4982,
+      "step": 470
+    },
+    {
+      "grad_norm": 0.7494113445281982,
+      "learning_rate": 9.58e-05,
+      "loss": 0.4806,
+      "step": 480
+    },
+    {
+      "grad_norm": 0.7424038648605347,
+      "learning_rate": 9.78e-05,
+      "loss": 0.4712,
+      "step": 490
+    },
+    {
+      "grad_norm": 0.872270405292511,
+      "learning_rate": 9.98e-05,
+      "loss": 0.4461,
+      "step": 500
+    },
+    {
+      "grad_norm": 0.8901939392089844,
+      "learning_rate": 9.9999778549206e-05,
+      "loss": 0.4498,
+      "step": 510
+    },
+    {
+      "grad_norm": 0.7232180237770081,
+      "learning_rate": 9.999901304280685e-05,
+      "loss": 0.4327,
+      "step": 520
+    },
+    {
+      "grad_norm": 0.871911883354187,
+      "learning_rate": 9.999770075521164e-05,
+      "loss": 0.4471,
+      "step": 530
+    },
+    {
+      "grad_norm": 0.7424588203430176,
+      "learning_rate": 9.99958417007713e-05,
+      "loss": 0.4488,
+      "step": 540
+    },
+    {
+      "grad_norm": 0.7698591351509094,
+      "learning_rate": 9.999343589981615e-05,
+      "loss": 0.4227,
+      "step": 550
+    },
+    {
+      "grad_norm": 0.803632915019989,
+      "learning_rate": 9.999048337865568e-05,
+      "loss": 0.4237,
+      "step": 560
+    },
+    {
+      "grad_norm": 0.7364802360534668,
+      "learning_rate": 9.998698416957815e-05,
+      "loss": 0.4246,
+      "step": 570
+    },
+    {
+      "grad_norm": 0.773603081703186,
+      "learning_rate": 9.998293831085037e-05,
+      "loss": 0.416,
+      "step": 580
+    },
+    {
+      "grad_norm": 0.8321452736854553,
+      "learning_rate": 9.997834584671719e-05,
+      "loss": 0.3952,
+      "step": 590
+    },
+    {
+      "grad_norm": 0.7838306427001953,
+      "learning_rate": 9.997320682740107e-05,
+      "loss": 0.3968,
+      "step": 600
+    },
+    {
+      "grad_norm": 0.6822958588600159,
+      "learning_rate": 9.996752130910149e-05,
+      "loss": 0.398,
+      "step": 610
+    },
+    {
+      "grad_norm": 0.7033558487892151,
+      "learning_rate": 9.99612893539944e-05,
+      "loss": 0.3889,
+      "step": 620
+    },
+    {
+      "grad_norm": 0.7827883362770081,
+      "learning_rate": 9.995451103023144e-05,
+      "loss": 0.3728,
+      "step": 630
+    },
+    {
+      "grad_norm": 0.7544336318969727,
+      "learning_rate": 9.994718641193928e-05,
+      "loss": 0.359,
+      "step": 640
+    },
+    {
+      "grad_norm": 0.7719122171401978,
+      "learning_rate": 9.993931557921874e-05,
+      "loss": 0.3604,
+      "step": 650
+    },
+    {
+      "grad_norm": 0.8056475520133972,
+      "learning_rate": 9.993089861814402e-05,
+      "loss": 0.3521,
+      "step": 660
+    },
+    {
+      "grad_norm": 0.7582818269729614,
+      "learning_rate": 9.992193562076166e-05,
+      "loss": 0.3306,
+      "step": 670
+    },
+    {
+      "grad_norm": 0.6936190724372864,
+      "learning_rate": 9.991242668508954e-05,
+      "loss": 0.3423,
+      "step": 680
+    },
+    {
+      "grad_norm": 0.6538547277450562,
+      "learning_rate": 9.990237191511587e-05,
+      "loss": 0.3223,
+      "step": 690
+    },
+    {
+      "grad_norm": 0.6155351996421814,
+      "learning_rate": 9.989177142079802e-05,
+      "loss": 0.3221,
+      "step": 700
+    },
+    {
+      "grad_norm": 0.6279995441436768,
+      "learning_rate": 9.988062531806126e-05,
+      "loss": 0.3278,
+      "step": 710
+    },
+    {
+      "grad_norm": 0.6754355430603027,
+      "learning_rate": 9.986893372879762e-05,
+      "loss": 0.3246,
+      "step": 720
+    },
+    {
+      "grad_norm": 0.7451342940330505,
+      "learning_rate": 9.985669678086443e-05,
+      "loss": 0.3289,
+      "step": 730
+    },
+    {
+      "grad_norm": 0.7065511345863342,
+      "learning_rate": 9.984391460808298e-05,
+      "loss": 0.3212,
+      "step": 740
+    },
+    {
+      "grad_norm": 0.7068234086036682,
+      "learning_rate": 9.983058735023709e-05,
+      "loss": 0.3177,
+      "step": 750
+    },
+    {
+      "grad_norm": 0.7593755722045898,
+      "learning_rate": 9.98167151530715e-05,
+      "loss": 0.3045,
+      "step": 760
+    },
+    {
+      "grad_norm": 0.8361629843711853,
+      "learning_rate": 9.980229816829034e-05,
+      "loss": 0.3129,
+      "step": 770
+    },
+    {
+      "grad_norm": 0.7489441633224487,
+      "learning_rate": 9.978733655355544e-05,
+      "loss": 0.31,
+      "step": 780
+    },
+    {
+      "grad_norm": 0.6411837935447693,
+      "learning_rate": 9.977183047248464e-05,
+      "loss": 0.3069,
+      "step": 790
+    },
+    {
+      "grad_norm": 0.7056382298469543,
+      "learning_rate": 9.975578009464992e-05,
+      "loss": 0.3129,
+      "step": 800
+    },
+    {
+      "grad_norm": 0.7930185198783875,
+      "learning_rate": 9.97391855955757e-05,
+      "loss": 0.298,
+      "step": 810
+    },
+    {
+      "grad_norm": 0.599122941493988,
+      "learning_rate": 9.972204715673669e-05,
+      "loss": 0.3146,
+      "step": 820
+    },
+    {
+      "grad_norm": 0.6534635424613953,
+      "learning_rate": 9.970436496555617e-05,
+      "loss": 0.3018,
+      "step": 830
+    },
+    {
+      "grad_norm": 0.6447542309761047,
+      "learning_rate": 9.968613921540373e-05,
+      "loss": 0.2948,
+      "step": 840
+    },
+    {
+      "grad_norm": 0.6775692701339722,
+      "learning_rate": 9.966737010559326e-05,
+      "loss": 0.3,
+      "step": 850
+    },
+    {
+      "grad_norm": 0.6695685982704163,
+      "learning_rate": 9.964805784138072e-05,
+      "loss": 0.3019,
+      "step": 860
+    },
+    {
+      "grad_norm": 0.6188393235206604,
+      "learning_rate": 9.962820263396195e-05,
+      "loss": 0.2946,
+      "step": 870
+    },
+    {
+      "grad_norm": 0.7127240300178528,
+      "learning_rate": 9.960780470047033e-05,
+      "loss": 0.2991,
+      "step": 880
+    },
+    {
+      "grad_norm": 0.7711222171783447,
+      "learning_rate": 9.958686426397437e-05,
+      "loss": 0.3037,
+      "step": 890
+    },
+    {
+      "grad_norm": 0.7105098962783813,
+      "learning_rate": 9.956538155347534e-05,
+      "loss": 0.3018,
+      "step": 900
+    },
+    {
+      "grad_norm": 0.741014838218689,
+      "learning_rate": 9.95433568039047e-05,
+      "loss": 0.2962,
+      "step": 910
+    },
+    {
+      "grad_norm": 0.5895913243293762,
+      "learning_rate": 9.952079025612162e-05,
+      "loss": 0.3044,
+      "step": 920
+    },
+    {
+      "grad_norm": 0.6453182697296143,
+      "learning_rate": 9.949768215691022e-05,
+      "loss": 0.3162,
+      "step": 930
+    },
+    {
+      "grad_norm": 0.6810097098350525,
+      "learning_rate": 9.9474032758977e-05,
+      "loss": 0.2927,
+      "step": 940
+    },
+    {
+      "grad_norm": 0.5451614856719971,
+      "learning_rate": 9.944984232094794e-05,
+      "loss": 0.2884,
+      "step": 950
+    },
+    {
+      "grad_norm": 0.7031158804893494,
+      "learning_rate": 9.942511110736584e-05,
+      "loss": 0.296,
+      "step": 960
+    },
+    {
+      "grad_norm": 0.6098790168762207,
+      "learning_rate": 9.939983938868726e-05,
+      "loss": 0.2891,
+      "step": 970
+    },
+    {
+      "grad_norm": 0.6627298593521118,
+      "learning_rate": 9.93740274412797e-05,
+      "loss": 0.297,
+      "step": 980
+    },
+    {
+      "grad_norm": 0.6396430134773254,
+      "learning_rate": 9.934767554741846e-05,
+      "loss": 0.285,
+      "step": 990
+    },
+    {
+      "grad_norm": 0.7423224449157715,
+      "learning_rate": 9.932078399528361e-05,
+      "loss": 0.2794,
+      "step": 1000
+    },
+    {
+      "grad_norm": 0.7008031010627747,
+      "learning_rate": 9.929335307895689e-05,
+      "loss": 0.2899,
+      "step": 1010
+    },
+    {
+      "grad_norm": 0.8223061561584473,
+      "learning_rate": 9.926538309841839e-05,
+      "loss": 0.2756,
+      "step": 1020
+    },
+    {
+      "grad_norm": 0.5336363911628723,
+      "learning_rate": 9.923687435954334e-05,
+      "loss": 0.2775,
+      "step": 1030
+    },
+    {
+      "grad_norm": 0.6405630707740784,
+      "learning_rate": 9.920782717409873e-05,
+      "loss": 0.2672,
+      "step": 1040
+    },
+    {
+      "grad_norm": 0.777278482913971,
+      "learning_rate": 9.917824185973994e-05,
+      "loss": 0.2775,
+      "step": 1050
+    },
+    {
+      "grad_norm": 0.5885688662528992,
+      "learning_rate": 9.914811874000723e-05,
+      "loss": 0.2833,
+      "step": 1060
+    },
+    {
+      "grad_norm": 0.6537587642669678,
+      "learning_rate": 9.911745814432218e-05,
+      "loss": 0.2662,
+      "step": 1070
+    },
+    {
+      "grad_norm": 0.5877810120582581,
+      "learning_rate": 9.90862604079842e-05,
+      "loss": 0.2666,
+      "step": 1080
+    },
+    {
+      "grad_norm": 0.6117822527885437,
+      "learning_rate": 9.90545258721667e-05,
+      "loss": 0.2683,
+      "step": 1090
+    },
+    {
+      "grad_norm": 0.5072397589683533,
+      "learning_rate": 9.90222548839135e-05,
+      "loss": 0.2617,
+      "step": 1100
+    },
+    {
+      "grad_norm": 0.6845521330833435,
+      "learning_rate": 9.898944779613495e-05,
+      "loss": 0.2657,
+      "step": 1110
+    },
+    {
+      "grad_norm": 0.5944507718086243,
+      "learning_rate": 9.89561049676041e-05,
+      "loss": 0.2619,
+      "step": 1120
+    },
+    {
+      "grad_norm": 0.6640071272850037,
+      "learning_rate": 9.89222267629528e-05,
+      "loss": 0.264,
+      "step": 1130
+    },
+    {
+      "grad_norm": 0.6789794564247131,
+      "learning_rate": 9.888781355266763e-05,
+      "loss": 0.2568,
+      "step": 1140
+    },
+    {
+      "grad_norm": 0.6515581011772156,
+      "learning_rate": 9.885286571308598e-05,
+      "loss": 0.2679,
+      "step": 1150
+    },
+    {
+      "grad_norm": 0.5297350883483887,
+      "learning_rate": 9.881738362639182e-05,
+      "loss": 0.2603,
+      "step": 1160
+    },
+    {
+      "grad_norm": 0.7514210343360901,
+      "learning_rate": 9.878136768061154e-05,
+      "loss": 0.2687,
+      "step": 1170
+    },
+    {
+      "grad_norm": 0.6085397005081177,
+      "learning_rate": 9.874481826960979e-05,
+      "loss": 0.2582,
+      "step": 1180
+    },
+    {
+      "grad_norm": 0.7021992802619934,
+      "learning_rate": 9.870773579308503e-05,
+      "loss": 0.2608,
+      "step": 1190
+    },
+    {
+      "grad_norm": 0.5477655529975891,
+      "learning_rate": 9.867012065656533e-05,
+      "loss": 0.262,
+      "step": 1200
+    },
+    {
+      "grad_norm": 0.7221733927726746,
+      "learning_rate": 9.863197327140376e-05,
+      "loss": 0.2606,
+      "step": 1210
+    },
+    {
+      "grad_norm": 0.6812732219696045,
+      "learning_rate": 9.859329405477403e-05,
+      "loss": 0.2539,
+      "step": 1220
+    },
+    {
+      "grad_norm": 0.7430665493011475,
+      "learning_rate": 9.855408342966585e-05,
+      "loss": 0.2463,
+      "step": 1230
+    },
+    {
+      "grad_norm": 0.7184039950370789,
+      "learning_rate": 9.851434182488033e-05,
+      "loss": 0.2588,
+      "step": 1240
+    },
+    {
+      "grad_norm": 0.6903015971183777,
+      "learning_rate": 9.84740696750253e-05,
+      "loss": 0.2424,
+      "step": 1250
+    },
+    {
+      "grad_norm": 0.4971630275249481,
+      "learning_rate": 9.843326742051055e-05,
+      "loss": 0.2413,
+      "step": 1260
+    },
+    {
+      "grad_norm": 0.6412865519523621,
+      "learning_rate": 9.839193550754297e-05,
+      "loss": 0.2452,
+      "step": 1270
+    },
+    {
+      "grad_norm": 0.5874801278114319,
+      "learning_rate": 9.835007438812177e-05,
+      "loss": 0.2515,
+      "step": 1280
+    },
+    {
+      "grad_norm": 0.6570281386375427,
+      "learning_rate": 9.830768452003341e-05,
+      "loss": 0.2416,
+      "step": 1290
+    },
+    {
+      "grad_norm": 0.6285550594329834,
+      "learning_rate": 9.826476636684671e-05,
+      "loss": 0.2428,
+      "step": 1300
+    },
+    {
+      "grad_norm": 0.5869091749191284,
+      "learning_rate": 9.822132039790773e-05,
+      "loss": 0.2377,
+      "step": 1310
+    },
+    {
+      "grad_norm": 0.6836414933204651,
+      "learning_rate": 9.817734708833461e-05,
+      "loss": 0.2494,
+      "step": 1320
+    },
+    {
+      "grad_norm": 0.5906325578689575,
+      "learning_rate": 9.813284691901243e-05,
+      "loss": 0.2524,
+      "step": 1330
+    },
+    {
+      "grad_norm": 0.6981961727142334,
+      "learning_rate": 9.808782037658792e-05,
+      "loss": 0.2553,
+      "step": 1340
+    },
+    {
+      "grad_norm": 0.5506542921066284,
+      "learning_rate": 9.804226795346411e-05,
+      "loss": 0.2341,
+      "step": 1350
+    },
+    {
+      "grad_norm": 0.7154520750045776,
+      "learning_rate": 9.799619014779503e-05,
+      "loss": 0.2507,
+      "step": 1360
+    },
+    {
+      "grad_norm": 0.5752660632133484,
+      "learning_rate": 9.794958746348013e-05,
+      "loss": 0.239,
+      "step": 1370
+    },
+    {
+      "grad_norm": 0.6976874470710754,
+      "learning_rate": 9.790246041015896e-05,
+      "loss": 0.2413,
+      "step": 1380
+    },
+    {
+      "grad_norm": 0.6117538213729858,
+      "learning_rate": 9.785480950320538e-05,
+      "loss": 0.2359,
+      "step": 1390
+    },
+    {
+      "grad_norm": 0.8666089773178101,
+      "learning_rate": 9.78066352637221e-05,
+      "loss": 0.2319,
+      "step": 1400
+    },
+    {
+      "grad_norm": 0.619526743888855,
+      "learning_rate": 9.775793821853488e-05,
+      "loss": 0.2373,
+      "step": 1410
+    },
+    {
+      "grad_norm": 0.6452130675315857,
+      "learning_rate": 9.77087189001868e-05,
+      "loss": 0.2278,
+      "step": 1420
+    },
+    {
+      "grad_norm": 0.6076776385307312,
+      "learning_rate": 9.765897784693243e-05,
+      "loss": 0.2294,
+      "step": 1430
+    },
+    {
+      "grad_norm": 0.53803551197052,
+      "learning_rate": 9.760871560273197e-05,
+      "loss": 0.2167,
+      "step": 1440
+    },
+    {
+      "grad_norm": 0.5496801733970642,
+      "learning_rate": 9.755793271724526e-05,
+      "loss": 0.221,
+      "step": 1450
+    },
+    {
+      "grad_norm": 0.47394850850105286,
+      "learning_rate": 9.750662974582584e-05,
+      "loss": 0.2169,
+      "step": 1460
+    },
+    {
+      "grad_norm": 0.6206640601158142,
+      "learning_rate": 9.745480724951473e-05,
+      "loss": 0.2124,
+      "step": 1470
+    },
+    {
+      "grad_norm": 0.596524178981781,
+      "learning_rate": 9.740246579503447e-05,
+      "loss": 0.2207,
+      "step": 1480
+    },
+    {
+      "grad_norm": 0.5237082839012146,
+      "learning_rate": 9.734960595478284e-05,
+      "loss": 0.2171,
+      "step": 1490
+    },
+    {
+      "grad_norm": 0.7235731482505798,
+      "learning_rate": 9.729622830682657e-05,
+      "loss": 0.2248,
+      "step": 1500
+    },
+    {
+      "grad_norm": 0.6276888251304626,
+      "learning_rate": 9.724233343489504e-05,
+      "loss": 0.2105,
+      "step": 1510
+    },
+    {
+      "grad_norm": 0.6757756471633911,
+      "learning_rate": 9.718792192837396e-05,
+      "loss": 0.2133,
+      "step": 1520
+    },
+    {
+      "grad_norm": 0.7439154982566833,
+      "learning_rate": 9.713299438229886e-05,
+      "loss": 0.2105,
+      "step": 1530
+    },
+    {
+      "grad_norm": 0.655325710773468,
+      "learning_rate": 9.707755139734855e-05,
+      "loss": 0.1995,
+      "step": 1540
+    },
+    {
+      "grad_norm": 0.649254560470581,
+      "learning_rate": 9.702159357983866e-05,
+      "loss": 0.2108,
+      "step": 1550
+    },
+    {
+      "grad_norm": 0.5804458856582642,
+      "learning_rate": 9.696512154171492e-05,
+      "loss": 0.191,
+      "step": 1560
+    },
+    {
+      "grad_norm": 0.5513803362846375,
+      "learning_rate": 9.690813590054645e-05,
+      "loss": 0.1974,
+      "step": 1570
+    },
+    {
+      "grad_norm": 0.47522595524787903,
+      "learning_rate": 9.685063727951914e-05,
+      "loss": 0.1936,
+      "step": 1580
+    },
+    {
+      "grad_norm": 0.6138707399368286,
+      "learning_rate": 9.679262630742865e-05,
+      "loss": 0.1969,
+      "step": 1590
+    },
+    {
+      "grad_norm": 0.6096352934837341,
+      "learning_rate": 9.673410361867373e-05,
+      "loss": 0.2015,
+      "step": 1600
+    },
+    {
+      "grad_norm": 0.5980384349822998,
+      "learning_rate": 9.667506985324909e-05,
+      "loss": 0.2018,
+      "step": 1610
+    },
+    {
+      "grad_norm": 0.5127062201499939,
+      "learning_rate": 9.661552565673855e-05,
+      "loss": 0.1993,
+      "step": 1620
+    },
+    {
+      "grad_norm": 0.7627597451210022,
+      "learning_rate": 9.655547168030789e-05,
+      "loss": 0.1983,
+      "step": 1630
+    },
+    {
+      "grad_norm": 0.49055951833724976,
+      "learning_rate": 9.649490858069777e-05,
+      "loss": 0.1855,
+      "step": 1640
+    },
+    {
+      "grad_norm": 0.5626708269119263,
+      "learning_rate": 9.643383702021658e-05,
+      "loss": 0.1849,
+      "step": 1650
+    },
+    {
+      "grad_norm": 0.681830644607544,
+      "learning_rate": 9.637225766673307e-05,
+      "loss": 0.1987,
+      "step": 1660
+    },
+    {
+      "grad_norm": 0.5874575972557068,
+      "learning_rate": 9.631017119366922e-05,
+      "loss": 0.1833,
+      "step": 1670
+    },
+    {
+      "grad_norm": 0.67975914478302,
+      "learning_rate": 9.624757827999273e-05,
+      "loss": 0.1853,
+      "step": 1680
+    },
+    {
+      "grad_norm": 0.771556556224823,
+      "learning_rate": 9.618447961020971e-05,
+      "loss": 0.1949,
+      "step": 1690
+    },
+    {
+      "grad_norm": 0.4923052191734314,
+      "learning_rate": 9.612087587435707e-05,
+      "loss": 0.1762,
+      "step": 1700
+    },
+    {
+      "grad_norm": 0.5661659240722656,
+      "learning_rate": 9.605676776799508e-05,
+      "loss": 0.1694,
+      "step": 1710
+    },
+    {
+      "grad_norm": 0.6437504291534424,
+      "learning_rate": 9.599215599219973e-05,
+      "loss": 0.1787,
+      "step": 1720
+    },
+    {
+      "grad_norm": 0.5951746106147766,
+      "learning_rate": 9.592704125355505e-05,
+      "loss": 0.1727,
+      "step": 1730
+    },
+    {
+      "grad_norm": 0.5490140914916992,
+      "learning_rate": 9.586142426414538e-05,
+      "loss": 0.1796,
+      "step": 1740
+    },
+    {
+      "grad_norm": 0.5185776352882385,
+      "learning_rate": 9.57953057415476e-05,
+      "loss": 0.1653,
+      "step": 1750
+    },
+    {
+      "grad_norm": 0.5233441591262817,
+      "learning_rate": 9.572868640882328e-05,
+      "loss": 0.1676,
+      "step": 1760
+    },
+    {
+      "grad_norm": 0.5588975548744202,
+      "learning_rate": 9.56615669945108e-05,
+      "loss": 0.1705,
+      "step": 1770
+    },
+    {
+      "grad_norm": 0.6676847338676453,
+      "learning_rate": 9.55939482326173e-05,
+      "loss": 0.1631,
+      "step": 1780
+    },
+    {
+      "grad_norm": 0.5605248212814331,
+      "learning_rate": 9.552583086261069e-05,
+      "loss": 0.1756,
+      "step": 1790
+    },
+    {
+      "grad_norm": 0.6252167820930481,
+      "learning_rate": 9.545721562941168e-05,
+      "loss": 0.1787,
+      "step": 1800
+    },
+    {
+      "grad_norm": 0.6381475925445557,
+      "learning_rate": 9.538810328338543e-05,
+      "loss": 0.1645,
+      "step": 1810
+    },
+    {
+      "grad_norm": 0.5540987849235535,
+      "learning_rate": 9.531849458033349e-05,
+      "loss": 0.1642,
+      "step": 1820
+    },
+    {
+      "grad_norm": 0.5993743538856506,
+      "learning_rate": 9.524839028148547e-05,
+      "loss": 0.1678,
+      "step": 1830
+    },
+    {
+      "grad_norm": 0.6065798401832581,
+      "learning_rate": 9.517779115349077e-05,
+      "loss": 0.1691,
+      "step": 1840
+    },
+    {
+      "grad_norm": 0.700529158115387,
+      "learning_rate": 9.510669796841014e-05,
+      "loss": 0.1662,
+      "step": 1850
+    },
+    {
+      "grad_norm": 0.5384991765022278,
+      "learning_rate": 9.503511150370727e-05,
+      "loss": 0.1636,
+      "step": 1860
+    },
+    {
+      "grad_norm": 0.6855055689811707,
+      "learning_rate": 9.496303254224024e-05,
+      "loss": 0.1575,
+      "step": 1870
+    },
+    {
+      "grad_norm": 0.5889959931373596,
+      "learning_rate": 9.489046187225306e-05,
+      "loss": 0.1583,
+      "step": 1880
+    },
+    {
+      "grad_norm": 0.8633908629417419,
+      "learning_rate": 9.481740028736692e-05,
+      "loss": 0.1632,
+      "step": 1890
+    },
+    {
+      "grad_norm": 0.6984632611274719,
+      "learning_rate": 9.474384858657164e-05,
+      "loss": 0.1673,
+      "step": 1900
+    },
+    {
+      "grad_norm": 0.5713339447975159,
+      "learning_rate": 9.466980757421679e-05,
+      "loss": 0.147,
+      "step": 1910
+    },
+    {
+      "grad_norm": 0.44927650690078735,
+      "learning_rate": 9.459527806000305e-05,
+      "loss": 0.1422,
+      "step": 1920
+    },
+    {
+      "grad_norm": 0.6698815226554871,
+      "learning_rate": 9.452026085897325e-05,
+      "loss": 0.1565,
+      "step": 1930
+    },
+    {
+      "grad_norm": 0.562569260597229,
+      "learning_rate": 9.444475679150348e-05,
+      "loss": 0.1388,
+      "step": 1940
+    },
+    {
+      "grad_norm": 0.6383896470069885,
+      "learning_rate": 9.436876668329411e-05,
+      "loss": 0.1452,
+      "step": 1950
+    },
+    {
+      "grad_norm": 0.7231951355934143,
+      "learning_rate": 9.429229136536079e-05,
+      "loss": 0.1518,
+      "step": 1960
+    },
+    {
+      "grad_norm": 0.6307895183563232,
+      "learning_rate": 9.421533167402534e-05,
+      "loss": 0.1496,
+      "step": 1970
+    },
+    {
+      "grad_norm": 0.550251841545105,
+      "learning_rate": 9.413788845090666e-05,
+      "loss": 0.1483,
+      "step": 1980
+    },
+    {
+      "grad_norm": 0.464999794960022,
+      "learning_rate": 9.405996254291136e-05,
+      "loss": 0.14,
+      "step": 1990
+    },
+    {
+      "grad_norm": 0.7153506278991699,
+      "learning_rate": 9.398155480222474e-05,
+      "loss": 0.1495,
+      "step": 2000
+    },
+    {
+      "grad_norm": 0.6995857954025269,
+      "learning_rate": 9.390266608630128e-05,
+      "loss": 0.1479,
+      "step": 2010
+    },
+    {
+      "grad_norm": 0.7008767127990723,
+      "learning_rate": 9.38232972578553e-05,
+      "loss": 0.1425,
+      "step": 2020
+    },
+    {
+      "grad_norm": 0.6410026550292969,
+      "learning_rate": 9.374344918485164e-05,
+      "loss": 0.1434,
+      "step": 2030
+    },
+    {
+      "grad_norm": 0.5032914280891418,
+      "learning_rate": 9.366312274049602e-05,
+      "loss": 0.1403,
+      "step": 2040
+    },
+    {
+      "grad_norm": 0.5063714385032654,
+      "learning_rate": 9.358231880322554e-05,
+      "loss": 0.134,
+      "step": 2050
+    },
+    {
+      "grad_norm": 0.6793342232704163,
+      "learning_rate": 9.350103825669916e-05,
+      "loss": 0.1375,
+      "step": 2060
+    },
+    {
+      "grad_norm": 0.8683373928070068,
+      "learning_rate": 9.341928198978787e-05,
+      "loss": 0.13,
+      "step": 2070
+    },
+    {
+      "grad_norm": 0.7127571702003479,
+      "learning_rate": 9.333705089656512e-05,
+      "loss": 0.1367,
+      "step": 2080
+    },
+    {
+      "grad_norm": 0.4944964349269867,
+      "learning_rate": 9.325434587629698e-05,
+      "loss": 0.1229,
+      "step": 2090
+    },
+    {
+      "grad_norm": 0.46358051896095276,
+      "learning_rate": 9.31711678334323e-05,
+      "loss": 0.123,
+      "step": 2100
+    },
+    {
+      "grad_norm": 0.6384886503219604,
+      "learning_rate": 9.308751767759282e-05,
+      "loss": 0.1294,
+      "step": 2110
+    },
+    {
+      "grad_norm": 0.7091969847679138,
+      "learning_rate": 9.300339632356325e-05,
+      "loss": 0.1351,
+      "step": 2120
+    },
+    {
+      "grad_norm": 0.6336053013801575,
+      "learning_rate": 9.291880469128124e-05,
+      "loss": 0.1344,
+      "step": 2130
+    },
+    {
+      "grad_norm": 0.5317314863204956,
+      "learning_rate": 9.283374370582732e-05,
+      "loss": 0.1196,
+      "step": 2140
+    },
+    {
+      "grad_norm": 0.483270525932312,
+      "learning_rate": 9.274821429741482e-05,
+      "loss": 0.1186,
+      "step": 2150
+    },
+    {
+      "grad_norm": 0.7253044247627258,
+      "learning_rate": 9.266221740137961e-05,
+      "loss": 0.1312,
+      "step": 2160
+    },
+    {
+      "grad_norm": 0.46905243396759033,
+      "learning_rate": 9.257575395817001e-05,
+      "loss": 0.1107,
+      "step": 2170
+    },
+    {
+      "grad_norm": 0.6425495743751526,
+      "learning_rate": 9.248882491333637e-05,
+      "loss": 0.104,
+      "step": 2180
+    },
+    {
+      "grad_norm": 0.6437727212905884,
+      "learning_rate": 9.240143121752076e-05,
+      "loss": 0.1123,
+      "step": 2190
+    },
+    {
+      "grad_norm": 0.5737977623939514,
+      "learning_rate": 9.23135738264467e-05,
+      "loss": 0.1067,
+      "step": 2200
+    },
+    {
+      "grad_norm": 0.5645055770874023,
+      "learning_rate": 9.222525370090849e-05,
+      "loss": 0.1132,
+      "step": 2210
+    },
+    {
+      "grad_norm": 0.7823241949081421,
+      "learning_rate": 9.213647180676088e-05,
+      "loss": 0.1139,
+      "step": 2220
+    },
+    {
+      "grad_norm": 0.4880165457725525,
+      "learning_rate": 9.204722911490846e-05,
+      "loss": 0.113,
+      "step": 2230
+    },
+    {
+      "grad_norm": 0.5553488731384277,
+      "learning_rate": 9.1957526601295e-05,
+      "loss": 0.1062,
+      "step": 2240
+    },
+    {
+      "grad_norm": 0.6223841309547424,
+      "learning_rate": 9.186736524689281e-05,
+      "loss": 0.1093,
+      "step": 2250
+    },
+    {
+      "grad_norm": 0.5920901894569397,
+      "learning_rate": 9.177674603769204e-05,
+      "loss": 0.1153,
+      "step": 2260
+    },
+    {
+      "grad_norm": 0.7499790191650391,
+      "learning_rate": 9.168566996468983e-05,
+      "loss": 0.1138,
+      "step": 2270
+    },
+    {
+      "grad_norm": 0.6586080193519592,
+      "learning_rate": 9.159413802387951e-05,
+      "loss": 0.1081,
+      "step": 2280
+    },
+    {
+      "grad_norm": 0.6859294772148132,
+      "learning_rate": 9.150215121623974e-05,
+      "loss": 0.1105,
+      "step": 2290
+    },
+    {
+      "grad_norm": 0.5580191016197205,
+      "learning_rate": 9.140971054772349e-05,
+      "loss": 0.1058,
+      "step": 2300
+    },
+    {
+      "grad_norm": 0.6616307497024536,
+      "learning_rate": 9.131681702924713e-05,
+      "loss": 0.105,
+      "step": 2310
+    },
+    {
+      "grad_norm": 0.6552684307098389,
+      "learning_rate": 9.122347167667926e-05,
+      "loss": 0.1093,
+      "step": 2320
+    },
+    {
+      "grad_norm": 0.4829959273338318,
+      "learning_rate": 9.112967551082973e-05,
+      "loss": 0.1032,
+      "step": 2330
+    },
+    {
+      "grad_norm": 0.5543748140335083,
+      "learning_rate": 9.103542955743835e-05,
+      "loss": 0.1024,
+      "step": 2340
+    },
+    {
+      "grad_norm": 0.7374470233917236,
+      "learning_rate": 9.094073484716381e-05,
+      "loss": 0.1132,
+      "step": 2350
+    },
+    {
+      "grad_norm": 0.5558682084083557,
+      "learning_rate": 9.084559241557226e-05,
+      "loss": 0.0989,
+      "step": 2360
+    },
+    {
+      "grad_norm": 0.5707993507385254,
+      "learning_rate": 9.075000330312608e-05,
+      "loss": 0.1084,
+      "step": 2370
+    },
+    {
+      "grad_norm": 0.5783806443214417,
+      "learning_rate": 9.065396855517253e-05,
+      "loss": 0.1011,
+      "step": 2380
+    },
+    {
+      "grad_norm": 0.6275078058242798,
+      "learning_rate": 9.055748922193219e-05,
+      "loss": 0.1144,
+      "step": 2390
+    },
+    {
+      "grad_norm": 0.5171684622764587,
+      "learning_rate": 9.046056635848761e-05,
+      "loss": 0.0974,
+      "step": 2400
+    },
+    {
+      "grad_norm": 0.5880749225616455,
+      "learning_rate": 9.036320102477169e-05,
+      "loss": 0.1014,
+      "step": 2410
+    },
+    {
+      "grad_norm": 0.5729674696922302,
+      "learning_rate": 9.02653942855561e-05,
+      "loss": 0.1014,
+      "step": 2420
+    },
+    {
+      "grad_norm": 0.5982695817947388,
+      "learning_rate": 9.016714721043971e-05,
+      "loss": 0.0967,
+      "step": 2430
+    },
+    {
+      "grad_norm": 0.61464524269104,
+      "learning_rate": 9.006846087383675e-05,
+      "loss": 0.0997,
+      "step": 2440
+    },
+    {
+      "grad_norm": 0.6253776550292969,
+      "learning_rate": 8.996933635496523e-05,
+      "loss": 0.1046,
+      "step": 2450
+    },
+    {
+      "grad_norm": 0.565126359462738,
+      "learning_rate": 8.986977473783498e-05,
+      "loss": 0.1,
+      "step": 2460
+    },
+    {
+      "grad_norm": 0.537251889705658,
+      "learning_rate": 8.97697771112359e-05,
+      "loss": 0.0985,
+      "step": 2470
+    },
+    {
+      "grad_norm": 0.5415845513343811,
+      "learning_rate": 8.966934456872602e-05,
+      "loss": 0.0948,
+      "step": 2480
+    },
+    {
+      "grad_norm": 0.568098783493042,
+      "learning_rate": 8.95684782086195e-05,
+      "loss": 0.1059,
+      "step": 2490
+    },
+    {
+      "grad_norm": 0.6046519875526428,
+      "learning_rate": 8.946717913397476e-05,
+      "loss": 0.0994,
+      "step": 2500
+    },
+    {
+      "grad_norm": 0.5644616484642029,
+      "learning_rate": 8.93654484525822e-05,
+      "loss": 0.0935,
+      "step": 2510
+    },
+    {
+      "grad_norm": 0.46616819500923157,
+      "learning_rate": 8.926328727695226e-05,
+      "loss": 0.0875,
+      "step": 2520
+    },
+    {
+      "grad_norm": 0.5794873833656311,
+      "learning_rate": 8.916069672430319e-05,
+      "loss": 0.0899,
+      "step": 2530
+    },
+    {
+      "grad_norm": 0.534477949142456,
+      "learning_rate": 8.905767791654884e-05,
+      "loss": 0.1025,
+      "step": 2540
+    },
+    {
+      "grad_norm": 0.5642363429069519,
+      "learning_rate": 8.895423198028638e-05,
+      "loss": 0.085,
+      "step": 2550
+    },
+    {
+      "grad_norm": 1.099249243736267,
+      "learning_rate": 8.885036004678402e-05,
+      "loss": 0.0958,
+      "step": 2560
+    },
+    {
+      "grad_norm": 0.6737452149391174,
+      "learning_rate": 8.874606325196857e-05,
+      "loss": 0.098,
+      "step": 2570
+    },
+    {
+      "grad_norm": 0.6400195956230164,
+      "learning_rate": 8.864134273641304e-05,
+      "loss": 0.0922,
+      "step": 2580
+    },
+    {
+      "grad_norm": 0.5365176200866699,
+      "learning_rate": 8.853619964532427e-05,
+      "loss": 0.0794,
+      "step": 2590
+    },
+    {
+      "grad_norm": 0.6816346645355225,
+      "learning_rate": 8.843063512853019e-05,
+      "loss": 0.0891,
+      "step": 2600
+    },
+    {
+      "grad_norm": 0.6123414635658264,
+      "learning_rate": 8.832465034046749e-05,
+      "loss": 0.084,
+      "step": 2610
+    },
+    {
+      "grad_norm": 0.557112455368042,
+      "learning_rate": 8.821824644016882e-05,
+      "loss": 0.0741,
+      "step": 2620
+    },
+    {
+      "grad_norm": 0.5661392211914062,
+      "learning_rate": 8.811142459125019e-05,
+      "loss": 0.0812,
+      "step": 2630
+    },
+    {
+      "grad_norm": 0.5407488346099854,
+      "learning_rate": 8.800418596189822e-05,
+      "loss": 0.0804,
+      "step": 2640
+    },
+    {
+      "grad_norm": 0.49320241808891296,
+      "learning_rate": 8.789653172485737e-05,
+      "loss": 0.0769,
+      "step": 2650
+    },
+    {
+      "grad_norm": 0.6032452583312988,
+      "learning_rate": 8.778846305741715e-05,
+      "loss": 0.0728,
+      "step": 2660
+    },
+    {
+      "grad_norm": 0.4812912344932556,
+      "learning_rate": 8.767998114139918e-05,
+      "loss": 0.0789,
+      "step": 2670
+    },
+    {
+      "grad_norm": 0.6122419238090515,
+      "learning_rate": 8.757108716314429e-05,
+      "loss": 0.0757,
+      "step": 2680
+    },
+    {
+      "grad_norm": 0.5506284236907959,
+      "learning_rate": 8.746178231349962e-05,
+      "loss": 0.0689,
+      "step": 2690
+    },
+    {
+      "grad_norm": 0.508624792098999,
+      "learning_rate": 8.735206778780549e-05,
+      "loss": 0.073,
+      "step": 2700
+    },
+    {
+      "grad_norm": 0.768593430519104,
+      "learning_rate": 8.724194478588234e-05,
+      "loss": 0.0746,
+      "step": 2710
+    },
+    {
+      "grad_norm": 0.5582564473152161,
+      "learning_rate": 8.713141451201772e-05,
+      "loss": 0.0825,
+      "step": 2720
+    },
+    {
+      "grad_norm": 0.47975367307662964,
+      "learning_rate": 8.702047817495295e-05,
+      "loss": 0.0662,
+      "step": 2730
+    },
+    {
+      "grad_norm": 0.4449773132801056,
+      "learning_rate": 8.69091369878701e-05,
+      "loss": 0.0705,
+      "step": 2740
+    },
+    {
+      "grad_norm": 0.6592329144477844,
+      "learning_rate": 8.679739216837849e-05,
+      "loss": 0.07,
+      "step": 2750
+    },
+    {
+      "grad_norm": 0.5570034980773926,
+      "learning_rate": 8.66852449385016e-05,
+      "loss": 0.0731,
+      "step": 2760
+    },
+    {
+      "grad_norm": 0.4954206347465515,
+      "learning_rate": 8.657269652466356e-05,
+      "loss": 0.073,
+      "step": 2770
+    },
+    {
+      "grad_norm": 0.5347866415977478,
+      "learning_rate": 8.645974815767577e-05,
+      "loss": 0.0692,
+      "step": 2780
+    },
+    {
+      "grad_norm": 0.6434792280197144,
+      "learning_rate": 8.634640107272351e-05,
+      "loss": 0.0759,
+      "step": 2790
+    },
+    {
+      "grad_norm": 0.5991449952125549,
+      "learning_rate": 8.623265650935234e-05,
+      "loss": 0.0725,
+      "step": 2800
+    },
+    {
+      "grad_norm": 0.49437442421913147,
+      "learning_rate": 8.611851571145456e-05,
+      "loss": 0.0653,
+      "step": 2810
+    },
+    {
+      "grad_norm": 0.482238233089447,
+      "learning_rate": 8.600397992725566e-05,
+      "loss": 0.0604,
+      "step": 2820
+    },
+    {
+      "grad_norm": 0.723780632019043,
+      "learning_rate": 8.588905040930061e-05,
+      "loss": 0.0739,
+      "step": 2830
+    },
+    {
+      "grad_norm": 0.6436156630516052,
+      "learning_rate": 8.577372841444022e-05,
+      "loss": 0.0642,
+      "step": 2840
+    },
+    {
+      "grad_norm": 0.667643129825592,
+      "learning_rate": 8.565801520381736e-05,
+      "loss": 0.0639,
+      "step": 2850
+    },
+    {
+      "grad_norm": 0.832385241985321,
+      "learning_rate": 8.554191204285313e-05,
+      "loss": 0.0655,
+      "step": 2860
+    },
+    {
+      "grad_norm": 0.7529436945915222,
+      "learning_rate": 8.542542020123315e-05,
+      "loss": 0.0627,
+      "step": 2870
+    },
+    {
+      "grad_norm": 0.537646472454071,
+      "learning_rate": 8.530854095289347e-05,
+      "loss": 0.0546,
+      "step": 2880
+    },
+    {
+      "grad_norm": 0.5955092310905457,
+      "learning_rate": 8.519127557600688e-05,
+      "loss": 0.0615,
+      "step": 2890
+    },
+    {
+      "grad_norm": 0.5789425373077393,
+      "learning_rate": 8.507362535296871e-05,
+      "loss": 0.06,
+      "step": 2900
+    },
+    {
+      "grad_norm": 0.5650882720947266,
+      "learning_rate": 8.495559157038299e-05,
+      "loss": 0.0596,
+      "step": 2910
+    },
+    {
+      "grad_norm": 0.432746022939682,
+      "learning_rate": 8.483717551904823e-05,
+      "loss": 0.0591,
+      "step": 2920
+    },
+    {
+      "grad_norm": 0.515415608882904,
+      "learning_rate": 8.47183784939434e-05,
+      "loss": 0.0546,
+      "step": 2930
+    },
+    {
+      "grad_norm": 0.5641840696334839,
+      "learning_rate": 8.459920179421374e-05,
+      "loss": 0.0604,
+      "step": 2940
+    },
+    {
+      "grad_norm": 0.6555214524269104,
+      "learning_rate": 8.447964672315656e-05,
+      "loss": 0.0648,
+      "step": 2950
+    },
+    {
+      "grad_norm": 0.567559540271759,
+      "learning_rate": 8.435971458820692e-05,
+      "loss": 0.0568,
+      "step": 2960
+    },
+    {
+      "grad_norm": 0.6113653182983398,
+      "learning_rate": 8.423940670092345e-05,
+      "loss": 0.0573,
+      "step": 2970
+    },
+    {
+      "grad_norm": 0.5120463967323303,
+      "learning_rate": 8.411872437697394e-05,
+      "loss": 0.0505,
+      "step": 2980
+    },
+    {
+      "grad_norm": 0.5301066040992737,
+      "learning_rate": 8.399766893612096e-05,
+      "loss": 0.0574,
+      "step": 2990
+    },
+    {
+      "grad_norm": 0.5676479339599609,
+      "learning_rate": 8.38762417022074e-05,
+      "loss": 0.0672,
+      "step": 3000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 10000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9223372036854775807,
+  "save_steps": 3000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

gr00t/run_3_real/checkpoint-3000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad069f2136738ad1d5be32c92812831e4eab7f84f59a54047d3d72150c0d6ae7
+size 7697

gr00t/run_3_real/checkpoint-3000/wandb_config.json

@@ -0,0 +1 @@
+{"project": "finetune-gr00t-n1d6", "run_id": "run_3_real"}

gr00t/run_3_real/checkpoint-3000/zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device, weights_only=False)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if ZERO_STAGE not in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size
+
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info("Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info("Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

gr00t/run_3_real/checkpoint-6000/config.json

@@ -0,0 +1,70 @@
+{
+  "action_horizon": 50,
+  "add_pos_embed": true,
+  "apply_sincos_state_encoding": true,
+  "architectures": [
+    "Gr00tN1d6"
+  ],
+  "attn_dropout": 0.2,
+  "attn_implementation": null,
+  "backbone_embedding_dim": 2048,
+  "backbone_model_type": "eagle",
+  "backbone_trainable_params_fp32": true,
+  "collator_overwrite_image_inputs": false,
+  "color_jitter_params": {
+    "brightness": 0.1,
+    "contrast": 0.1,
+    "hue": 0.1,
+    "saturation": 0.1
+  },
+  "crop_fraction": 0.95,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 48,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "interleave_self_attention": true,
+    "norm_type": "ada_norm",
+    "num_attention_heads": 32,
+    "num_layers": 32,
+    "output_dim": 1024,
+    "positional_embeddings": null
+  },
+  "eagle_collator": true,
+  "formalize_language": true,
+  "gemma_collator": false,
+  "hidden_size": 1024,
+  "image_crop_size": null,
+  "image_target_size": null,
+  "input_embedding_dim": 1536,
+  "load_bf16": true,
+  "max_action_dim": 128,
+  "max_num_embodiments": 32,
+  "max_seq_len": 1024,
+  "max_state_dim": 128,
+  "model_dtype": "bfloat16",
+  "model_name": "nvidia/Eagle-Block2A-2B-v2",
+  "model_type": "Gr00tN1d6",
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_inference_timesteps": 4,
+  "num_timestep_buckets": 1000,
+  "random_rotation_angle": null,
+  "reproject_vision": false,
+  "select_layer": 16,
+  "shortest_image_edge": 256,
+  "state_dropout_prob": 0.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "tune_diffusion_model": true,
+  "tune_llm": true,
+  "tune_projector": true,
+  "tune_top_llm_layers": 4,
+  "tune_visual": true,
+  "tune_vlln": true,
+  "use_albumentations_transforms": true,
+  "use_alternate_vl_dit": true,
+  "use_flash_attention": true,
+  "use_relative_action": true,
+  "use_vlln": true
+}

gr00t/run_3_real/checkpoint-6000/embodiment_id.json

@@ -0,0 +1,11 @@
+{
+  "robocasa_panda_omron": 13,
+  "gr1": 20,
+  "behavior_r1_pro": 24,
+  "unitree_g1": 8,
+  "oxe_google": 0,
+  "oxe_widowx": 1,
+  "libero_panda": 2,
+  "oxe_droid": 16,
+  "new_embodiment": 10
+}

gr00t/run_3_real/checkpoint-6000/experiment_cfg/conf.yaml

@@ -0,0 +1,246 @@
+load_config_path: null
+model:
+  model_type: Gr00tN1d6
+  model_dtype: bfloat16
+  model_name: nvidia/Eagle-Block2A-2B-v2
+  backbone_model_type: eagle
+  model_revision: null
+  tune_top_llm_layers: 4
+  backbone_embedding_dim: 2048
+  tune_llm: true
+  tune_visual: true
+  select_layer: 16
+  reproject_vision: false
+  use_flash_attention: true
+  load_bf16: false
+  collator_overwrite_image_inputs: false
+  eagle_collator: true
+  backbone_trainable_params_fp32: true
+  image_crop_size: null
+  image_target_size: null
+  shortest_image_edge: 256
+  crop_fraction: 0.95
+  random_rotation_angle: null
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    saturation: 0.5
+    hue: 0.08
+  use_albumentations_transforms: true
+  formalize_language: true
+  apply_sincos_state_encoding: false
+  use_relative_action: true
+  max_state_dim: 29
+  max_action_dim: 29
+  action_horizon: 16
+  hidden_size: 1024
+  input_embedding_dim: 1536
+  add_pos_embed: true
+  attn_dropout: 0.2
+  use_vlln: true
+  max_seq_len: 1024
+  use_alternate_vl_dit: true
+  attend_text_every_n_blocks: 2
+  diffusion_model_cfg:
+    positional_embeddings: null
+    num_layers: 32
+    num_attention_heads: 32
+    attention_head_dim: 48
+    norm_type: ada_norm
+    dropout: 0.2
+    final_dropout: true
+    output_dim: 1024
+    interleave_self_attention: true
+  num_inference_timesteps: 4
+  noise_beta_alpha: 1.5
+  noise_beta_beta: 1.0
+  noise_s: 0.999
+  num_timestep_buckets: 1000
+  tune_projector: true
+  tune_diffusion_model: true
+  tune_vlln: true
+  state_dropout_prob: 0.0
+  state_additive_noise_scale: 0.0
+  max_num_embodiments: 32
+data:
+  datasets:
+  - dataset_paths:
+    - /workspace/capstone-vla/dataset_tools/data/realWorldPouring
+    embodiment_tag: new_embodiment
+    mix_ratio: 1.0
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  modality_configs:
+    new_embodiment:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - ego_view
+        - left_wrist_view
+        - right_wrist_view
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+  download_cache: false
+  shard_size: 1024
+  episode_sampling_rate: 0.1
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  mode: single_turn
+  random_chop: 0.0
+  mock_dataset_mode: false
+  shuffle: true
+  seed: 42
+  multiprocessing_context: fork
+  allow_padding: false
+  subsample_ratio: 1.0
+  image_crop_size:
+  - 244
+  - 244
+  image_target_size:
+  - 224
+  - 224
+  video_backend: torchcodec
+training:
+  output_dir: /workspace/capstone-vla/third_party/gr00t/chkpt_rlwd/run_3_real
+  experiment_name: null
+  max_steps: 10000
+  global_batch_size: 32
+  batch_size: null
+  gradient_accumulation_steps: 1
+  learning_rate: 0.0001
+  lr_scheduler_type: cosine
+  weight_decay: 1.0e-05
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  max_grad_norm: 1.0
+  optim: adamw_torch
+  start_from_checkpoint: nvidia/GR00T-N1.6-3B
+  tf32: true
+  fp16: false
+  bf16: true
+  eval_bf16: true
+  logging_steps: 10
+  save_steps: 3000
+  save_total_limit: 3
+  save_vl_model: false
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  max_concurrent_uploads: 2
+  eval_strategy: 'no'
+  eval_steps: 500
+  eval_set_split_ratio: 0.1
+  eval_batch_size: 2
+  save_best_eval_metric_name: ''
+  save_best_eval_metric_greater_is_better: true
+  deepspeed_stage: 2
+  gradient_checkpointing: false
+  transformers_trust_remote_code: true
+  transformers_local_files_only: false
+  transformers_cache_dir: null
+  transformers_access_token: null
+  use_ddp: false
+  ddp_bucket_cap_mb: 100
+  num_gpus: 2
+  dataloader_num_workers: 4
+  remove_unused_columns: false
+  use_wandb: true
+  wandb_project: finetune-gr00t-n1d6
+  enable_profiling: false
+  max_retries: 3
+  assert_loss_less_than: null
+  add_rl_callback: false
+  enable_open_loop_eval: false
+  open_loop_eval_traj_ids:
+  - 0
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_plot_indices: null
+max_steps: 10000
+save_steps: 3000

gr00t/run_3_real/checkpoint-6000/experiment_cfg/config.yaml

@@ -0,0 +1,277 @@
+!!python/object:gr00t.configs.base_config.Config
+data: !!python/object:gr00t.configs.data.data_config.DataConfig
+  allow_padding: false
+  datasets:
+  - !!python/object:gr00t.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /workspace/capstone-vla/dataset_tools/data/realWorldPouring
+    dataset_type: physical_embodiment
+    embodiment_tag: new_embodiment
+    mix_ratio: 1.0
+    val_dataset_path: null
+  download_cache: false
+  episode_sampling_rate: 0.1
+  image_crop_size:
+  - 244
+  - 244
+  image_target_size:
+  - 224
+  - 224
+  mock_dataset_mode: false
+  modality_configs:
+    new_embodiment:
+      action: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: &id001 !!python/object/apply:gr00t.data.types.ActionFormat
+          - default
+          rep: &id002 !!python/object/apply:gr00t.data.types.ActionRepresentation
+          - absolute
+          state_key: null
+          type: &id003 !!python/object/apply:gr00t.data.types.ActionType
+          - non_eef
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:gr00t.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys: &id004
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+      language: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys: *id004
+        sin_cos_embedding_keys: null
+      video: !!python/object:gr00t.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - ego_view
+        - left_wrist_view
+        - right_wrist_view
+        sin_cos_embedding_keys: null
+  mode: single_turn
+  multiprocessing_context: fork
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  random_chop: 0.0
+  seed: 42
+  shard_size: 1024
+  shuffle: true
+  subsample_ratio: 1.0
+  video_backend: torchcodec
+load_config_path: null
+model: !!python/object:gr00t.configs.model.gr00t_n1d6.Gr00tN1d6Config
+  _attn_implementation_autoset: false
+  _attn_implementation_internal: null
+  _commit_hash: null
+  _name_or_path: ''
+  add_cross_attention: false
+  architectures: null
+  backbone_model_type: eagle
+  backbone_trainable_params_fp32: true
+  bad_words_ids: null
+  begin_suppress_tokens: null
+  bos_token_id: null
+  chunk_size_feed_forward: 0
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    hue: 0.08
+    saturation: 0.5
+  cross_attention_hidden_size: null
+  decoder_start_token_id: null
+  diffusion_model_cfg:
+    attention_head_dim: 48
+    dropout: 0.2
+    final_dropout: true
+    interleave_self_attention: true
+    norm_type: ada_norm
+    num_attention_heads: 32
+    num_layers: 32
+    output_dim: 1024
+    positional_embeddings: null
+  diversity_penalty: 0.0
+  do_sample: false
+  eagle_collator: true
+  early_stopping: false
+  encoder_no_repeat_ngram_size: 0
+  eos_token_id: null
+  exponential_decay_length_penalty: null
+  finetuning_task: null
+  forced_bos_token_id: null
+  forced_eos_token_id: null
+  id2label:
+    0: LABEL_0
+    1: LABEL_1
+  is_decoder: false
+  is_encoder_decoder: false
+  label2id:
+    LABEL_0: 0
+    LABEL_1: 1
+  length_penalty: 1.0
+  load_bf16: false
+  max_length: 20
+  min_length: 0
+  model_name: nvidia/Eagle-Block2A-2B-v2
+  no_repeat_ngram_size: 0
+  num_beam_groups: 1
+  num_beams: 1
+  num_return_sequences: 1
+  output_attentions: false
+  output_hidden_states: false
+  output_scores: false
+  pad_token_id: null
+  prefix: null
+  problem_type: null
+  pruned_heads: {}
+  random_rotation_angle: null
+  remove_invalid_values: false
+  repetition_penalty: 1.0
+  reproject_vision: false
+  return_dict: true
+  return_dict_in_generate: false
+  sep_token_id: null
+  state_dropout_prob: 0.0
+  suppress_tokens: null
+  task_specific_params: null
+  temperature: 1.0
+  tf_legacy_loss: false
+  tie_encoder_decoder: false
+  tie_word_embeddings: true
+  tokenizer_class: null
+  top_k: 50
+  top_p: 1.0
+  torch_dtype: null
+  torchscript: false
+  transformers_version: null
+  tune_diffusion_model: true
+  tune_llm: true
+  tune_projector: true
+  tune_visual: true
+  typical_p: 1.0
+  use_bfloat16: false
+  use_relative_action: true
+training: !!python/object:gr00t.configs.training.training_config.TrainingConfig
+  add_rl_callback: false
+  assert_loss_less_than: null
+  batch_size: null
+  bf16: true
+  dataloader_num_workers: 4
+  ddp_bucket_cap_mb: 100
+  deepspeed_stage: 2
+  enable_open_loop_eval: false
+  enable_profiling: false
+  eval_batch_size: 2
+  eval_bf16: true
+  eval_set_split_ratio: 0.1
+  eval_steps: 500
+  eval_strategy: 'no'
+  experiment_name: null
+  fp16: false
+  global_batch_size: 32
+  gradient_accumulation_steps: 1
+  gradient_checkpointing: false
+  learning_rate: 0.0001
+  logging_steps: 10
+  lr_scheduler_type: cosine
+  max_concurrent_uploads: 2
+  max_grad_norm: 1.0
+  max_retries: 3
+  max_steps: 10000
+  num_gpus: 2
+  open_loop_eval_plot_indices: null
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_traj_ids:
+  - 0
+  optim: adamw_torch
+  output_dir: /workspace/capstone-vla/third_party/gr00t/chkpt_rlwd/run_3_real
+  remove_unused_columns: false
+  save_best_eval_metric_greater_is_better: true
+  save_best_eval_metric_name: ''
+  save_steps: 3000
+  save_total_limit: 3
+  save_vl_model: false
+  start_from_checkpoint: nvidia/GR00T-N1.6-3B
+  tf32: true
+  transformers_access_token: null
+  transformers_cache_dir: null
+  transformers_local_files_only: false
+  transformers_trust_remote_code: true
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  use_ddp: false
+  use_wandb: true
+  wandb_project: finetune-gr00t-n1d6
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  weight_decay: 1.0e-05

gr00t/run_3_real/checkpoint-6000/experiment_cfg/dataset_statistics.json

@@ -0,0 +1,761 @@
+{
+  "new_embodiment": {
+    "state": {
+      "left_arm": {
+        "min": [
+          -1.4764578342437744,
+          -0.1518283635377884,
+          -1.499962568283081,
+          -2.27626633644104,
+          -1.5671002864837646,
+          -0.6656755208969116,
+          -0.5207215547561646
+        ],
+        "max": [
+          0.3441833555698395,
+          0.657663881778717,
+          0.04959612339735031,
+          -0.7633329629898071,
+          0.5798436999320984,
+          0.6755949258804321,
+          0.6866571307182312
+        ],
+        "mean": [
+          -0.24634727835655212,
+          0.2629535496234894,
+          -0.5132859945297241,
+          -1.822871208190918,
+          -0.15705835819244385,
+          0.250943660736084,
+          -0.009801508858799934
+        ],
+        "std": [
+          0.2749035060405731,
+          0.061663031578063965,
+          0.23117874562740326,
+          0.2519250214099884,
+          0.42646297812461853,
+          0.20134037733078003,
+          0.23961548507213593
+        ],
+        "q01": [
+          -1.1976326751708983,
+          0.1270393806695938,
+          -1.415273575782776,
+          -2.1694675636291505,
+          -1.555278000831604,
+          -0.4085486590862274,
+          -0.3467614698410034
+        ],
+        "q99": [
+          0.11830186098813932,
+          0.4688370370864868,
+          -0.08621232867240906,
+          -1.1188685512542724,
+          0.2598086118698119,
+          0.6740685701370239,
+          0.5260765767097467
+        ]
+      },
+      "left_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.7819951176643372,
+          -0.15371035039424896,
+          -0.7819951176643372,
+          -0.7819951176643372,
+          -1.5639902353286743,
+          -1.5639902353286743
+        ],
+        "std": [
+          1.2748703956604004,
+          1.4928919076919556,
+          1.2748703956604004,
+          1.2748703956604004,
+          2.549740791320801,
+          2.549740791320801
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "left_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "neck": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "right_arm": {
+        "min": [
+          -0.7119174599647522,
+          -0.55275958776474,
+          -0.09536845237016678,
+          -2.144679546356201,
+          -0.6652951240539551,
+          -0.5672546625137329,
+          -0.5874713063240051
+        ],
+        "max": [
+          0.33939826488494873,
+          -0.002669741865247488,
+          0.928513765335083,
+          -0.8445846438407898,
+          1.433588981628418,
+          0.4833299517631531,
+          0.6599544286727905
+        ],
+        "mean": [
+          -0.21578556299209595,
+          -0.22310534119606018,
+          0.40323933959007263,
+          -1.6887274980545044,
+          0.08409612625837326,
+          -0.18936818838119507,
+          -0.007405608426779509
+        ],
+        "std": [
+          0.16777803003787994,
+          0.07015040516853333,
+          0.16941194236278534,
+          0.19395458698272705,
+          0.5333251953125,
+          0.1218024268746376,
+          0.20256562530994415
+        ],
+        "q01": [
+          -0.5456185221672059,
+          -0.4390794897079468,
+          0.054558104127645494,
+          -2.009228744506836,
+          -0.5162186360359191,
+          -0.4283971929550171,
+          -0.3479098403453827
+        ],
+        "q99": [
+          0.16945502102374999,
+          -0.08812180131673814,
+          0.7629516124725342,
+          -1.1490095472335817,
+          1.3221839523315426,
+          0.1579314386844635,
+          0.45433507800102235
+        ]
+      },
+      "right_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.41444578766822815,
+          0.15046590566635132,
+          -0.41444578766822815,
+          -0.41444578766822815,
+          -0.8288915753364563,
+          -0.8288915753364563
+        ],
+        "std": [
+          1.432570219039917,
+          1.4905503988265991,
+          1.432570219039917,
+          1.432570219039917,
+          2.865140438079834,
+          2.865140438079834
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "right_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "waist": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      }
+    },
+    "action": {
+      "left_arm": {
+        "min": [
+          -1.5419785976409912,
+          -0.19808706641197205,
+          -1.5707963705062866,
+          -2.3319737911224365,
+          -1.5707963705062866,
+          -0.6781747937202454,
+          -0.5912933349609375
+        ],
+        "max": [
+          0.339247465133667,
+          0.6958394646644592,
+          0.09750956296920776,
+          -0.7315306067466736,
+          0.6010454297065735,
+          0.6981316804885864,
+          0.6981316804885864
+        ],
+        "mean": [
+          -0.27577918767929077,
+          0.265458881855011,
+          -0.5064833760261536,
+          -1.8468353748321533,
+          -0.1631959080696106,
+          0.25038862228393555,
+          -0.011975681409239769
+        ],
+        "std": [
+          0.284289687871933,
+          0.06655233353376389,
+          0.23614443838596344,
+          0.26587340235710144,
+          0.432668536901474,
+          0.20466120541095734,
+          0.24292133748531342
+        ],
+        "q01": [
+          -1.254055471420288,
+          0.11745307803153991,
+          -1.4238929557800293,
+          -2.2081959438323975,
+          -1.5707963705062866,
+          -0.41540132522583006,
+          -0.3490602993965149
+        ],
+        "q99": [
+          0.09790542602539062,
+          0.48043997287750184,
+          -0.06798712402582203,
+          -1.1039583587646489,
+          0.2601294279098509,
+          0.6981316804885864,
+          0.53674639225006
+        ]
+      },
+      "left_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.7821466326713562,
+          -0.15372000634670258,
+          -0.7821466326713562,
+          -0.7821466326713562,
+          -1.5642932653427124,
+          -1.5642932653427124
+        ],
+        "std": [
+          1.2747758626937866,
+          1.4928909540176392,
+          1.2747758626937866,
+          1.2747758626937866,
+          2.5495517253875732,
+          2.5495517253875732
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "left_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "neck": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "right_arm": {
+        "min": [
+          -0.7609559297561646,
+          -0.5897848010063171,
+          -0.13661164045333862,
+          -2.1815736293792725,
+          -0.6809128522872925,
+          -0.5955591201782227,
+          -0.6164834499359131
+        ],
+        "max": [
+          0.31390857696533203,
+          0.01057685911655426,
+          0.9158902764320374,
+          -0.8264364004135132,
+          1.5061734914779663,
+          0.49332451820373535,
+          0.6635355353355408
+        ],
+        "mean": [
+          -0.24524866044521332,
+          -0.2248098999261856,
+          0.39028772711753845,
+          -1.7206827402114868,
+          0.08302607387304306,
+          -0.18993926048278809,
+          -0.012360971421003342
+        ],
+        "std": [
+          0.17419038712978363,
+          0.07439576834440231,
+          0.17066127061843872,
+          0.2032645046710968,
+          0.5400993824005127,
+          0.124636709690094,
+          0.20725084841251373
+        ],
+        "q01": [
+          -0.5854583382606506,
+          -0.45275834798812864,
+          0.03628820136189461,
+          -2.0545105361938476,
+          -0.5252219843864441,
+          -0.44349650502204896,
+          -0.36548688769340515
+        ],
+        "q99": [
+          0.15696598112583027,
+          -0.07897948294878006,
+          0.7490021729469298,
+          -1.1520879220962532,
+          1.3330879974365206,
+          0.15911160767078353,
+          0.4541356778144836
+        ]
+      },
+      "right_hand": {
+        "min": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "max": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ],
+        "mean": [
+          -0.414561003446579,
+          0.15042728185653687,
+          -0.414561003446579,
+          -0.414561003446579,
+          -0.829122006893158,
+          -0.829122006893158
+        ],
+        "std": [
+          1.4325326681137085,
+          1.4905540943145752,
+          1.4325326681137085,
+          1.4325326681137085,
+          2.865065336227417,
+          2.865065336227417
+        ],
+        "q01": [
+          -1.5,
+          -1.5,
+          -1.5,
+          -1.5,
+          -3.0,
+          -3.0
+        ],
+        "q99": [
+          1.5,
+          1.5,
+          1.5,
+          1.5,
+          3.0,
+          3.0
+        ]
+      },
+      "right_leg": {
+        "min": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0,
+          0.0
+        ]
+      },
+      "waist": {
+        "min": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "max": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "mean": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "std": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q01": [
+          0.0,
+          0.0,
+          0.0
+        ],
+        "q99": [
+          0.0,
+          0.0,
+          0.0
+        ]
+      }
+    },
+    "relative_action": {}
+  }
+}

gr00t/run_3_real/checkpoint-6000/experiment_cfg/final_model_config.json

@@ -0,0 +1,53 @@
+{
+  "model_type": "Gr00tN1d6",
+  "model_dtype": "bfloat16",
+  "model_name": "nvidia/Eagle-Block2A-2B-v2",
+  "backbone_model_type": "eagle",
+  "model_revision": null,
+  "tune_top_llm_layers": 4,
+  "backbone_embedding_dim": 2048,
+  "tune_llm": true,
+  "tune_visual": true,
+  "select_layer": 16,
+  "reproject_vision": false,
+  "use_flash_attention": true,
+  "load_bf16": true,
+  "collator_overwrite_image_inputs": false,
+  "eagle_collator": true,
+  "backbone_trainable_params_fp32": true,
+  "apply_sincos_state_encoding": true,
+  "use_relative_action": true,
+  "max_state_dim": 128,
+  "max_action_dim": 128,
+  "action_horizon": 50,
+  "hidden_size": 1024,
+  "input_embedding_dim": 1536,
+  "add_pos_embed": true,
+  "attn_dropout": 0.2,
+  "use_vlln": true,
+  "max_seq_len": 1024,
+  "use_alternate_vl_dit": true,
+  "attend_text_every_n_blocks": 2,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 48,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "interleave_self_attention": true,
+    "norm_type": "ada_norm",
+    "num_attention_heads": 32,
+    "num_layers": 32,
+    "output_dim": 1024,
+    "positional_embeddings": null
+  },
+  "num_inference_timesteps": 4,
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_timestep_buckets": 1000,
+  "tune_projector": true,
+  "tune_diffusion_model": true,
+  "tune_vlln": true,
+  "state_dropout_prob": 0.0,
+  "state_additive_noise_scale": 0.0,
+  "max_num_embodiments": 32
+}

gr00t/run_3_real/checkpoint-6000/global_step6000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70a7bc3d0cb005321f6eae4c18d4984bf0ee60d52fdab55aaa745bf882b350e2
+size 19719732513

gr00t/run_3_real/checkpoint-6000/global_step6000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b26315aea5b97c59a037b3cb47e8a4f3cdb696f72a720628836dc6d6069b057f
+size 19719703265

gr00t/run_3_real/checkpoint-6000/global_step6000/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca53f0d7a6dda6fdc5553d544d4d18f9e2d1f975def9f7894b053b681d55ddee
+size 6573616357

gr00t/run_3_real/checkpoint-6000/latest

@@ -0,0 +1 @@
+global_step6000

gr00t/run_3_real/checkpoint-6000/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ad36989ff2cb0ee1c8ded1ede6843ce5b936bcec8ef8463bce64f4d14d54865
+size 4991091456

gr00t/run_3_real/checkpoint-6000/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e001de8d57c950622598d10c5c121ad332a87129cb51a4b2d693f8ceb6b77281
+size 1582283096

gr00t/run_3_real/checkpoint-6000/processor_config.json

@@ -0,0 +1,504 @@
+{
+  "processor_class": "Gr00tN1d6Processor",
+  "processor_kwargs": {
+    "modality_configs": {
+      "behavior_r1_pro": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "observation.images.rgb.head_256_256",
+            "observation.images.rgb.left_wrist_256_256",
+            "observation.images.rgb.right_wrist_256_256"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "robot_pos",
+            "robot_ori_cos",
+            "robot_ori_sin",
+            "robot_2d_ori",
+            "robot_2d_ori_cos",
+            "robot_2d_ori_sin",
+            "robot_lin_vel",
+            "robot_ang_vel",
+            "arm_left_qpos",
+            "arm_left_qpos_sin",
+            "arm_left_qpos_cos",
+            "eef_left_pos",
+            "eef_left_quat",
+            "gripper_left_qpos",
+            "arm_right_qpos",
+            "arm_right_qpos_sin",
+            "arm_right_qpos_cos",
+            "eef_right_pos",
+            "eef_right_quat",
+            "gripper_right_qpos",
+            "trunk_qpos"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15,
+            16,
+            17,
+            18,
+            19,
+            20,
+            21,
+            22,
+            23,
+            24,
+            25,
+            26,
+            27,
+            28,
+            29,
+            30,
+            31
+          ],
+          "modality_keys": [
+            "base",
+            "torso",
+            "left_arm",
+            "left_gripper",
+            "right_arm",
+            "right_gripper"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "trunk_qpos"
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "arm_left_qpos"
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": "arm_right_qpos"
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.coarse_action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "gr1": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "ego_view_bg_crop_pad_res256_freq20"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "right_arm",
+            "left_hand",
+            "right_hand",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "RELATIVE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "task"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "robocasa_panda_omron": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "res256_image_side_0",
+            "res256_image_side_1",
+            "res256_image_wrist_0"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position_relative",
+            "end_effector_rotation_relative",
+            "gripper_qpos",
+            "base_position",
+            "base_rotation"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close",
+            "base_motion",
+            "control_mode"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "new_embodiment": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "ego_view",
+            "left_wrist_view",
+            "right_wrist_view"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      }
+    },
+    "image_crop_size": null,
+    "image_target_size": null,
+    "use_albumentations": true,
+    "random_rotation_angle": null,
+    "color_jitter_params": {
+      "brightness": 0.3,
+      "contrast": 0.4,
+      "saturation": 0.5,
+      "hue": 0.08
+    },
+    "shortest_image_edge": 256,
+    "crop_fraction": 0.95,
+    "model_name": "nvidia/Eagle-Block2A-2B-v2",
+    "model_type": "eagle",
+    "formalize_language": true,
+    "max_state_dim": 128,
+    "max_action_dim": 128,
+    "max_action_horizon": 50,
+    "use_percentiles": false,
+    "clip_outliers": true,
+    "apply_sincos_state_encoding": true,
+    "use_relative_action": true
+  }
+}

gr00t/run_3_real/checkpoint-6000/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c5d19f4a0a74e119ecf959246f583700d04e4d211554b2c5d70819c149684a5
+size 14917

gr00t/run_3_real/checkpoint-6000/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f5551718a1ab06ed579a68e98548d6e6456889d49ce75b8623c5c0b6a9c4269
+size 14917

gr00t/run_3_real/checkpoint-6000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c81168c968205dd4f18f99afb8d250e3039fa2276a497474189eec84cb61b541
+size 1465