Upload folder using huggingface_hub

checkpoint-200/embodiment_id.json

@@ -52,5 +52,6 @@
   "libero_sim": 2,
   "droid_sim": 3,
   "unitree_g1_sonic": 11,
-  "new_embodiment": 10
+  "new_embodiment": 10,
+  "robocasa_gr1_tabletop": 10
 }

checkpoint-200/experiment_cfg/conf.yaml

@@ -233,7 +233,7 @@ training:
   transformers_access_token: null
   use_ddp: false
   ddp_bucket_cap_mb: 100
-  num_gpus: 8
+  num_gpus: 4
   dataloader_num_workers: 6
   remove_unused_columns: false
   use_wandb: true

checkpoint-200/experiment_cfg/config.yaml

@@ -245,7 +245,7 @@ training: !!python/object:gr00t.configs.training.training_config.TrainingConfig
   max_grad_norm: 1.0
   max_retries: 3
   max_steps: 200
-  num_gpus: 8
+  num_gpus: 4
   open_loop_eval_plot_indices: null
   open_loop_eval_steps_per_traj: 100
   open_loop_eval_traj_ids:

checkpoint-200/experiment_cfg/final_processor_config.json

@@ -21,7 +21,7 @@
   "random_rotation_angle": "0",
   "color_jitter_params": "{'brightness': 0.3, 'contrast': 0.4, 'saturation': 0.5, 'hue': 0.08}",
   "processor": "Qwen3VLProcessor:\n- image_processor: Qwen2VLImageProcessorFast {\n  \"crop_size\": null,\n  \"data_format\": \"channels_first\",\n  \"default_to_square\": true,\n  \"device\": null,\n  \"disable_grouping\": null,\n  \"do_center_crop\": null,\n  \"do_convert_rgb\": true,\n  \"do_normalize\": true,\n  \"do_pad\": null,\n  \"do_rescale\": true,\n  \"do_resize\": true,\n  \"image_mean\": [\n    0.5,\n    0.5,\n    0.5\n  ],\n  \"image_processor_type\": \"Qwen2VLImageProcessorFast\",\n  \"image_std\": [\n    0.5,\n    0.5,\n    0.5\n  ],\n  \"input_data_format\": null,\n  \"max_pixels\": null,\n  \"merge_size\": 2,\n  \"min_pixels\": null,\n  \"pad_size\": null,\n  \"patch_size\": 16,\n  \"processor_class\": \"Qwen3VLProcessor\",\n  \"resample\": 3,\n  \"rescale_factor\": 0.00392156862745098,\n  \"return_tensors\": null,\n  \"size\": {\n    \"longest_edge\": 16777216,\n    \"shortest_edge\": 65536\n  },\n  \"temporal_patch_size\": 2\n}\n\n- tokenizer: Qwen2TokenizerFast(name_or_path='nvidia/Cosmos-Reason2-2B', vocab_size=151643, model_max_length=262144, is_fast=True, padding_side='left', truncation_side='right', special_tokens={'eos_token': '<|im_end|>', 'pad_token': '<|endoftext|>', 'additional_special_tokens': ['<|im_start|>', '<|im_end|>', '<|object_ref_start|>', '<|object_ref_end|>', '<|box_start|>', '<|box_end|>', '<|quad_start|>', '<|quad_end|>', '<|vision_start|>', '<|vision_end|>', '<|vision_pad|>', '<|image_pad|>', '<|video_pad|>']}, clean_up_tokenization_spaces=False, added_tokens_decoder={\n\t151643: AddedToken(\"<|endoftext|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151644: AddedToken(\"<|im_start|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151645: AddedToken(\"<|im_end|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151646: AddedToken(\"<|object_ref_start|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151647: AddedToken(\"<|object_ref_end|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151648: AddedToken(\"<|box_start|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151649: AddedToken(\"<|box_end|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151650: AddedToken(\"<|quad_start|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151651: AddedToken(\"<|quad_end|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151652: AddedToken(\"<|vision_start|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151653: AddedToken(\"<|vision_end|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151654: AddedToken(\"<|vision_pad|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151655: AddedToken(\"<|image_pad|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151656: AddedToken(\"<|video_pad|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),\n\t151657: AddedToken(\"<tool_call>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151658: AddedToken(\"</tool_call>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151659: AddedToken(\"<|fim_prefix|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151660: AddedToken(\"<|fim_middle|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151661: AddedToken(\"<|fim_suffix|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151662: AddedToken(\"<|fim_pad|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151663: AddedToken(\"<|repo_name|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151664: AddedToken(\"<|file_sep|>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151665: AddedToken(\"<tool_response>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151666: AddedToken(\"</tool_response>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151667: AddedToken(\"<think>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n\t151668: AddedToken(\"</think>\", rstrip=False, lstrip=False, single_word=False, normalized=False, special=False),\n}\n)\n- video_processor: Qwen3VLVideoProcessor {\n  \"crop_size\": null,\n  \"data_format\": \"channels_first\",\n  \"default_to_square\": true,\n  \"device\": null,\n  \"do_center_crop\": null,\n  \"do_convert_rgb\": true,\n  \"do_normalize\": true,\n  \"do_rescale\": true,\n  \"do_resize\": true,\n  \"do_sample_frames\": true,\n  \"fps\": 2,\n  \"image_mean\": [\n    0.5,\n    0.5,\n    0.5\n  ],\n  \"image_std\": [\n    0.5,\n    0.5,\n    0.5\n  ],\n  \"input_data_format\": null,\n  \"max_frames\": 768,\n  \"merge_size\": 2,\n  \"min_frames\": 4,\n  \"num_frames\": null,\n  \"pad_size\": null,\n  \"patch_size\": 16,\n  \"processor_class\": \"Qwen3VLProcessor\",\n  \"resample\": 3,\n  \"rescale_factor\": 0.00392156862745098,\n  \"return_metadata\": false,\n  \"size\": {\n    \"longest_edge\": 25165824,\n    \"shortest_edge\": 4096\n  },\n  \"temporal_patch_size\": 2,\n  \"video_metadata\": null,\n  \"video_processor_type\": \"Qwen3VLVideoProcessor\"\n}\n\n\n{\n  \"processor_class\": \"Qwen3VLProcessor\"\n}\n",
-  "embodiment_id_mapping": "{'robocasa_panda_omron': 13, 'oxe_droid': 17, 'oxe_fractal': 18, 'oxe_language_table': 19, 'oxe_bridge': 20, 'unknown': 22, 'gr1_unified': 20, 'agibot': 26, 'sim_behavior_r1_pro': 23, 'xdof': 24, 'xdof_oss_data': 25, 'unitree_g1_full_body_with_waist_height_nav_cmd': 25, 'real_r1_pro_sharpa': 27, 'real_r1_pro_sharpa_add_view': 27, 'real_r1_pro_sharpa_relative_arm_joint': 26, 'real_r1_pro_sharpa_delta_eef': 26, 'real_r1_pro_sharpa_absolute_eef': 26, 'real_r1_pro_sharpa_meanstd': 26, 'real_r1_pro_sharpa_relative_eef': 26, 'real_r1_pro_sharpa_relative_eef_add_view': 26, 'real_r1_pro_sharpa_relative_eef_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_human': 26, 'real_r1_pro_sharpa_relative_eef_human_add_view': 26, 'real_r1_pro_sharpa_relative_eef_human_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_egodex': 26, 'real_r1_pro_sharpa_relative_eef_egodex_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_egodex_wrist_only': 26, 'real_r1_pro_sharpa_relative_eef_maxinsights': 26, 'real_r1_pro_sharpa_relative_eef_maxinsights_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_mecka': 26, 'real_r1_pro_sharpa_relative_eef_mecka_relative_hand': 26, 'real_g1_relative_eef_absolute_joints': 25, 'real_g1_relative_eef_absolute_joints_wrist_cam': 25, 'real_g1_relative_eef_relative_joints': 25, 'real_r1_pro_sharpa_relative_eef_relative_hand_relative_joint': 26, 'real_r1_pro_sharpa_relative_joint': 29, 'oxe_droid_relative_eef_relative_joint': 24, 'oxe_droid_relative_eef_relative_joint_swapped': 24, 'oxe_droid_relative_eef_relative_joint_upweight_z': 24, 'oxe_droid_relative_eef_relative_joint_upweight_z_swapped': 24, 'oxe_droid_relative_eef_relative_joint_3view': 24, 'oxe_droid_relative_eef_relative_joint_3view_swapped': 24, 'oxe_droid_relative_eef': 24, 'oxe_droid_joint_position_relative': 24, 'xdof_relative_eef_relative_joint': 27, 'xdof_relative_eef_relative_joint_subtask': 27, 'xdof_relative_eef': 27, 'xdof_relative_joint': 28, 'simpler_env_google': 0, 'simpler_env_widowx': 1, 'libero_sim': 2, 'droid_sim': 3, 'unitree_g1_sonic': 11, 'new_embodiment': 10}",
+  "embodiment_id_mapping": "{'robocasa_panda_omron': 13, 'oxe_droid': 17, 'oxe_fractal': 18, 'oxe_language_table': 19, 'oxe_bridge': 20, 'unknown': 22, 'gr1_unified': 20, 'agibot': 26, 'sim_behavior_r1_pro': 23, 'xdof': 24, 'xdof_oss_data': 25, 'unitree_g1_full_body_with_waist_height_nav_cmd': 25, 'real_r1_pro_sharpa': 27, 'real_r1_pro_sharpa_add_view': 27, 'real_r1_pro_sharpa_relative_arm_joint': 26, 'real_r1_pro_sharpa_delta_eef': 26, 'real_r1_pro_sharpa_absolute_eef': 26, 'real_r1_pro_sharpa_meanstd': 26, 'real_r1_pro_sharpa_relative_eef': 26, 'real_r1_pro_sharpa_relative_eef_add_view': 26, 'real_r1_pro_sharpa_relative_eef_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_human': 26, 'real_r1_pro_sharpa_relative_eef_human_add_view': 26, 'real_r1_pro_sharpa_relative_eef_human_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_egodex': 26, 'real_r1_pro_sharpa_relative_eef_egodex_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_egodex_wrist_only': 26, 'real_r1_pro_sharpa_relative_eef_maxinsights': 26, 'real_r1_pro_sharpa_relative_eef_maxinsights_relative_hand': 26, 'real_r1_pro_sharpa_relative_eef_mecka': 26, 'real_r1_pro_sharpa_relative_eef_mecka_relative_hand': 26, 'real_g1_relative_eef_absolute_joints': 25, 'real_g1_relative_eef_absolute_joints_wrist_cam': 25, 'real_g1_relative_eef_relative_joints': 25, 'real_r1_pro_sharpa_relative_eef_relative_hand_relative_joint': 26, 'real_r1_pro_sharpa_relative_joint': 29, 'oxe_droid_relative_eef_relative_joint': 24, 'oxe_droid_relative_eef_relative_joint_swapped': 24, 'oxe_droid_relative_eef_relative_joint_upweight_z': 24, 'oxe_droid_relative_eef_relative_joint_upweight_z_swapped': 24, 'oxe_droid_relative_eef_relative_joint_3view': 24, 'oxe_droid_relative_eef_relative_joint_3view_swapped': 24, 'oxe_droid_relative_eef': 24, 'oxe_droid_joint_position_relative': 24, 'xdof_relative_eef_relative_joint': 27, 'xdof_relative_eef_relative_joint_subtask': 27, 'xdof_relative_eef': 27, 'xdof_relative_joint': 28, 'simpler_env_google': 0, 'simpler_env_widowx': 1, 'libero_sim': 2, 'droid_sim': 3, 'unitree_g1_sonic': 11, 'new_embodiment': 10, 'robocasa_gr1_tabletop': 10}",
   "shortest_image_edge": "256",
   "crop_fraction": "0.95",
   "statistics": "{}",

checkpoint-200/global_step200/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e021031a158a91169a0d390758a7708c788530f4c7019d22d385ec242b1fdb7d
-size 2430787077
+oid sha256:c7a2988b9429d26f834daf54126c2aec1d74875fffa8ec7018acacd1062531a8
+size 4861568625

checkpoint-200/global_step200/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e81223895ea21953c79579bc0680bc43c6162b1206aa2448e71c573f42588ea6
-size 2430786949
+oid sha256:33bea903a22668985d9940f3352ac71c8ed2b93dc18fdc4c89069ae80a0f490d
+size 4861568369

checkpoint-200/global_step200/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:389f1017c3ac9a2ccdbb77f1e8f5792414e1f9402a882030788feacccd3053b3
-size 2430786821
+oid sha256:883614873568ec558410c180e6686fcced8be4b959ba000192a2932896f9705a
+size 4861566321

checkpoint-200/global_step200/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dde280f67066fff9ba3424430412c2dd6f526eab9218b44990fba90591aa5bc8
-size 2430786949
+oid sha256:67324024e5bd44a4f24c5c92a32660df09152988c95688df5cbeb6a0f4396db9
+size 4861563121

checkpoint-200/global_step200/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8c51f6ba26b6c749233b7c74aa3cb5490d6575c7a3ebe1c4084d8d536f9eb45b
+oid sha256:32f823a47fe5ee3ea4a78f50f44fb842b14d50c1413fa70e3e9874af118d9666
 size 9335640879

checkpoint-200/model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:767a661ebc07bed2a1a59109c88edf99851d3e68a92fd094852706e54bb7b018
+oid sha256:780ab19548b8f3ec6892023bfb21264b0babe9df35e4fe495dc281c81f8d87c7
 size 4990519232

checkpoint-200/model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1bab4341246788d49c97ced1095298d5dca911514b48f204fa137297bf84fbf0
+oid sha256:9a61270b555246d86a0b0483cb54f31265ae1d4f4c57acd64817cd244bfc15a6
 size 1919980184

checkpoint-200/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1da79915866c81b03aa6f4ae320e9b0b3dd9e893a33c2edcaf99b8274f703a6c
-size 16389
+oid sha256:585f5e83ab2e47062abc72edf125d0506c7fd66dab9442016b9eb5677cfa6221
+size 15429

checkpoint-200/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eeeb41463cb0bca10c9abcc7ee94548e073986499fec2c859de993587fe9f331
-size 16389
+oid sha256:5854ff429ff758a2a53c029ba5bdd58cf6e6b042d55f9732c23e9d0270fff740
+size 15429

checkpoint-200/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca554ac9d33e69623dde66b3c447075f49c3757a2a8b2ec0e6dcd34055e70563
-size 16389
+oid sha256:8e18c420a8712f7aa4cbc826e57af75af90687a03861f07cca8f7b457cc73557
+size 15429

checkpoint-200/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f162e35a49f155ed8a516e90893c3036186841e0ba9d2cf787a140412282ee8d
-size 16389
+oid sha256:fe3f74e890f0cc772511802269c2817b638ba2496bab58391696237ec9e78198
+size 15429

checkpoint-200/trainer_state.json

@@ -10,123 +10,123 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "grad_norm": 0.38291576504707336,
+      "grad_norm": 0.23940400779247284,
       "learning_rate": 9e-05,
-      "loss": 1.1938,
+      "loss": 1.1936,
       "step": 10
     },
     {
-      "grad_norm": 0.35215654969215393,
+      "grad_norm": 0.6992996335029602,
       "learning_rate": 9.944739353007344e-05,
-      "loss": 1.1394,
+      "loss": 1.1512,
       "step": 20
     },
     {
-      "grad_norm": 0.2596343159675598,
+      "grad_norm": 0.2213364690542221,
       "learning_rate": 9.755282581475769e-05,
-      "loss": 1.1093,
+      "loss": 1.1256,
       "step": 30
     },
     {
-      "grad_norm": 0.2617182433605194,
+      "grad_norm": 0.2311399132013321,
       "learning_rate": 9.43611409721806e-05,
-      "loss": 1.1161,
+      "loss": 1.0988,
       "step": 40
     },
     {
-      "grad_norm": 0.29306936264038086,
+      "grad_norm": 0.23056164383888245,
       "learning_rate": 8.995939984474624e-05,
-      "loss": 1.1066,
+      "loss": 1.1029,
       "step": 50
     },
     {
-      "grad_norm": 0.1921951025724411,
+      "grad_norm": 0.28725796937942505,
       "learning_rate": 8.44676704559283e-05,
-      "loss": 1.1083,
+      "loss": 1.102,
       "step": 60
     },
     {
-      "grad_norm": 0.24627622961997986,
+      "grad_norm": 0.33148351311683655,
       "learning_rate": 7.803575286758364e-05,
-      "loss": 1.0941,
+      "loss": 1.0932,
       "step": 70
     },
     {
-      "grad_norm": 0.20044036209583282,
+      "grad_norm": 0.1882592737674713,
       "learning_rate": 7.083909302476453e-05,
-      "loss": 1.1003,
+      "loss": 1.0951,
       "step": 80
     },
     {
-      "grad_norm": 0.2700786888599396,
+      "grad_norm": 0.1874360740184784,
       "learning_rate": 6.307399704769099e-05,
-      "loss": 1.0953,
+      "loss": 1.093,
       "step": 90
     },
     {
-      "grad_norm": 0.21156367659568787,
+      "grad_norm": 0.20323705673217773,
       "learning_rate": 5.495227651252315e-05,
-      "loss": 1.0812,
+      "loss": 1.0979,
       "step": 100
     },
     {
-      "grad_norm": 0.5570153594017029,
+      "grad_norm": 0.2660678029060364,
       "learning_rate": 4.669547078371504e-05,
-      "loss": 1.0969,
+      "loss": 1.0988,
       "step": 110
     },
     {
-      "grad_norm": 0.2210264801979065,
+      "grad_norm": 0.18318524956703186,
       "learning_rate": 3.852880399766243e-05,
-      "loss": 1.0823,
+      "loss": 1.0875,
       "step": 120
     },
     {
-      "grad_norm": 0.3023281991481781,
+      "grad_norm": 0.16894328594207764,
       "learning_rate": 3.0675041535377405e-05,
-      "loss": 1.0895,
+      "loss": 1.0838,
       "step": 130
     },
     {
-      "grad_norm": 0.19830650091171265,
+      "grad_norm": 0.28183799982070923,
       "learning_rate": 2.3348413563600325e-05,
-      "loss": 1.0849,
+      "loss": 1.0836,
       "step": 140
     },
     {
-      "grad_norm": 0.20713168382644653,
+      "grad_norm": 0.1599818468093872,
       "learning_rate": 1.6748771394307585e-05,
-      "loss": 1.0754,
+      "loss": 1.0826,
       "step": 150
     },
     {
-      "grad_norm": 0.2407350242137909,
+      "grad_norm": 0.19995185732841492,
       "learning_rate": 1.1056136061894384e-05,
-      "loss": 1.0815,
+      "loss": 1.0854,
       "step": 160
     },
     {
-      "grad_norm": 0.33742642402648926,
+      "grad_norm": 0.1766791194677353,
       "learning_rate": 6.425787818636131e-06,
-      "loss": 1.0768,
+      "loss": 1.0848,
       "step": 170
     },
     {
-      "grad_norm": 0.18564505875110626,
+      "grad_norm": 0.1526567041873932,
       "learning_rate": 2.9840304941919415e-06,
-      "loss": 1.0749,
+      "loss": 1.0879,
       "step": 180
     },
     {
-      "grad_norm": 0.24459722638130188,
+      "grad_norm": 0.18048691749572754,
       "learning_rate": 8.247462563808817e-07,
-      "loss": 1.0779,
+      "loss": 1.0787,
       "step": 190
     },
     {
-      "grad_norm": 0.2062901258468628,
+      "grad_norm": 0.17072460055351257,
       "learning_rate": 6.834750376549792e-09,
-      "loss": 1.073,
+      "loss": 1.0785,
       "step": 200
     }
   ],
@@ -148,7 +148,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 4,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-200/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f12f96c7f87961dd8b0507908f81c748e9819934b946731946bb687ab2f11d33
+oid sha256:8948bc0acba8971f43e26c05e378b8f61c5f795118c5779a55a4b9bba2bbeb33
 size 7825