Upload folder using huggingface_hub

ParaVLA-LIBERO/config.json

@@ -0,0 +1,105 @@
+{
+  "type": "paravla",
+  "n_obs_steps": 1,
+  "input_features": {
+    "observation.images.image": {
+      "type": "VISUAL",
+      "shape": [
+        3,
+        256,
+        256
+      ]
+    },
+    "observation.images.image2": {
+      "type": "VISUAL",
+      "shape": [
+        3,
+        256,
+        256
+      ]
+    },
+    "observation.state": {
+      "type": "STATE",
+      "shape": [
+        8
+      ]
+    }
+  },
+  "output_features": {
+    "action": {
+      "type": "ACTION",
+      "shape": [
+        7
+      ]
+    }
+  },
+  "device": "cuda",
+  "use_amp": true,
+  "use_peft": false,
+  "push_to_hub": false,
+  "repo_id": "GSR/ParaVLA-LIBERO",
+  "private": null,
+  "tags": null,
+  "license": null,
+  "pretrained_path": null,
+  "chunk_size": 16,
+  "n_action_steps": 8,
+  "use_temporal_ensembling": true,
+  "temporal_ensemble_coeff": 0.9,
+  "max_state_dim": 32,
+  "max_action_dim": 32,
+  "normalization_mapping": {
+    "VISUAL": "IDENTITY",
+    "STATE": "MEAN_STD",
+    "ACTION": "MEAN_STD"
+  },
+  "vision_model": "facebook/dinov2-large",
+  "text_model": "google-t5/t5-large",
+  "image_size": 224,
+  "hidden_size": 1024,
+  "fusion_heads": 16,
+  "fusion_layers": 4,
+  "dropout": 0.1,
+  "num_denoise_steps": 20,
+  "num_expert_layers": 8,
+  "num_expert_heads": 16,
+  "expert_head_dim": 64,
+  "expert_injection_stride": 1,
+  "expert_injection_min_scale": 1.0,
+  "use_coord_injection": false,
+  "freeze_vision_encoder": false,
+  "freeze_text_encoder": true,
+  "freeze_patch_embed": false,
+  "unfreeze_blocks": 4,
+  "use_lora": false,
+  "lora_r": 64,
+  "lora_alpha": 128,
+  "lora_dropout": 0.05,
+  "lora_target_modules": [
+    "qkv",
+    "proj",
+    "fc1",
+    "fc2",
+    "patch_embed.proj"
+  ],
+  "tokenizer_max_length": 64,
+  "text_dropout_prob": 0.0,
+  "vision_dim": 1536,
+  "use_goal_predictor": false,
+  "use_wrist_camera": true,
+  "guidance_scale": 1.0,
+  "film_scale_init": 1.0,
+  "optimizer_lr": 0.0001,
+  "backbone_lr_fraction": 0.1,
+  "patch_embed_lr_fraction": 0.5,
+  "optimizer_betas": [
+    0.9,
+    0.95
+  ],
+  "optimizer_eps": 1e-08,
+  "optimizer_weight_decay": 0.0005,
+  "optimizer_grad_clip_norm": 2.0,
+  "scheduler_warmup_steps": 2000,
+  "scheduler_decay_steps": 45000,
+  "scheduler_decay_lr": 1e-06
+}

ParaVLA-LIBERO/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06b84d29368c56a35f0e025d67bc537a786a3573a169cbf0373e76fa78ed37a3
+size 3946084960

ParaVLA-LIBERO/policy_postprocessor.json

@@ -0,0 +1,32 @@
+{
+  "name": "policy_postprocessor",
+  "steps": [
+    {
+      "registry_name": "unnormalizer_processor",
+      "config": {
+        "eps": 1e-08,
+        "features": {
+          "action": {
+            "type": "ACTION",
+            "shape": [
+              7
+            ]
+          }
+        },
+        "norm_map": {
+          "VISUAL": "IDENTITY",
+          "STATE": "MEAN_STD",
+          "ACTION": "MEAN_STD"
+        }
+      },
+      "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
+    },
+    {
+      "registry_name": "device_processor",
+      "config": {
+        "device": "cpu",
+        "float_dtype": null
+      }
+    }
+  ]
+}

ParaVLA-LIBERO/policy_postprocessor_step_0_unnormalizer_processor.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a002c0df7f79c5b169c5a899ad151d4ea1bed246c7d82bd93ed1556558d517a9
+size 7668

ParaVLA-LIBERO/policy_preprocessor.json

@@ -0,0 +1,85 @@
+{
+  "name": "policy_preprocessor",
+  "steps": [
+    {
+      "registry_name": "rename_observations_processor",
+      "config": {
+        "rename_map": {}
+      }
+    },
+    {
+      "registry_name": "to_batch_processor",
+      "config": {}
+    },
+    {
+      "registry_name": "paravla_default_task_processor",
+      "config": {}
+    },
+    {
+      "registry_name": "tokenizer_processor",
+      "config": {
+        "max_length": 64,
+        "task_key": "task",
+        "padding_side": "right",
+        "padding": "max_length",
+        "truncation": true,
+        "tokenizer_name": "t5-large"
+      }
+    },
+    {
+      "registry_name": "paravla_image_processor",
+      "config": {
+        "target_size": 224
+      }
+    },
+    {
+      "registry_name": "device_processor",
+      "config": {
+        "device": "cuda",
+        "float_dtype": null
+      }
+    },
+    {
+      "registry_name": "normalizer_processor",
+      "config": {
+        "eps": 1e-08,
+        "features": {
+          "observation.images.image": {
+            "type": "VISUAL",
+            "shape": [
+              3,
+              256,
+              256
+            ]
+          },
+          "observation.images.image2": {
+            "type": "VISUAL",
+            "shape": [
+              3,
+              256,
+              256
+            ]
+          },
+          "observation.state": {
+            "type": "STATE",
+            "shape": [
+              8
+            ]
+          },
+          "action": {
+            "type": "ACTION",
+            "shape": [
+              7
+            ]
+          }
+        },
+        "norm_map": {
+          "VISUAL": "IDENTITY",
+          "STATE": "MEAN_STD",
+          "ACTION": "MEAN_STD"
+        }
+      },
+      "state_file": "policy_preprocessor_step_6_normalizer_processor.safetensors"
+    }
+  ]
+}

ParaVLA-LIBERO/policy_preprocessor_step_6_normalizer_processor.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f98234b21ce8ffacf1b86657f3cf8e24ac6c31ebcd6a3b83cecd484439674306
+size 7676

ParaVLA-LIBERO/train_config.json

@@ -0,0 +1,342 @@
+{
+  "dataset": {
+    "repo_id": "HuggingFaceVLA/libero",
+    "root": null,
+    "episodes": null,
+    "image_transforms": {
+      "enable": true,
+      "max_num_transforms": 3,
+      "random_order": false,
+      "tfs": {
+        "brightness": {
+          "weight": 1.0,
+          "type": "ColorJitter",
+          "kwargs": {
+            "brightness": [
+              0.8,
+              1.2
+            ]
+          }
+        },
+        "contrast": {
+          "weight": 1.0,
+          "type": "ColorJitter",
+          "kwargs": {
+            "contrast": [
+              0.8,
+              1.2
+            ]
+          }
+        },
+        "saturation": {
+          "weight": 1.0,
+          "type": "ColorJitter",
+          "kwargs": {
+            "saturation": [
+              0.5,
+              1.5
+            ]
+          }
+        },
+        "hue": {
+          "weight": 1.0,
+          "type": "ColorJitter",
+          "kwargs": {
+            "hue": [
+              -0.05,
+              0.05
+            ]
+          }
+        },
+        "sharpness": {
+          "weight": 1.0,
+          "type": "SharpnessJitter",
+          "kwargs": {
+            "sharpness": [
+              0.5,
+              1.5
+            ]
+          }
+        },
+        "affine": {
+          "weight": 1.0,
+          "type": "RandomAffine",
+          "kwargs": {
+            "degrees": [
+              -5.0,
+              5.0
+            ],
+            "translate": [
+              0.05,
+              0.05
+            ]
+          }
+        }
+      }
+    },
+    "revision": null,
+    "use_imagenet_stats": true,
+    "video_backend": "torchcodec",
+    "return_uint8": false,
+    "streaming": false
+  },
+  "env": {
+    "type": "libero",
+    "task": "libero_spatial,libero_object,libero_goal,libero_10",
+    "fps": 30,
+    "features": {
+      "action": {
+        "type": "ACTION",
+        "shape": [
+          7
+        ]
+      },
+      "pixels/agentview_image": {
+        "type": "VISUAL",
+        "shape": [
+          360,
+          360,
+          3
+        ]
+      },
+      "pixels/robot0_eye_in_hand_image": {
+        "type": "VISUAL",
+        "shape": [
+          360,
+          360,
+          3
+        ]
+      },
+      "robot_state/eef/pos": {
+        "type": "STATE",
+        "shape": [
+          3
+        ]
+      },
+      "robot_state/eef/quat": {
+        "type": "STATE",
+        "shape": [
+          4
+        ]
+      },
+      "robot_state/eef/mat": {
+        "type": "STATE",
+        "shape": [
+          3,
+          3
+        ]
+      },
+      "robot_state/gripper/qpos": {
+        "type": "STATE",
+        "shape": [
+          2
+        ]
+      },
+      "robot_state/gripper/qvel": {
+        "type": "STATE",
+        "shape": [
+          2
+        ]
+      },
+      "robot_state/joints/pos": {
+        "type": "STATE",
+        "shape": [
+          7
+        ]
+      },
+      "robot_state/joints/vel": {
+        "type": "STATE",
+        "shape": [
+          7
+        ]
+      }
+    },
+    "features_map": {
+      "action": "action",
+      "robot_state/eef/pos": "observation.state.eef_pos",
+      "robot_state/eef/quat": "observation.state.eef_quat",
+      "robot_state/eef/mat": "observation.state.eef_mat",
+      "robot_state/gripper/qpos": "observation.state.gripper_qpos",
+      "robot_state/gripper/qvel": "observation.state.gripper_qvel",
+      "robot_state/joints/pos": "observation.state.joint_pos",
+      "robot_state/joints/vel": "observation.state.joint_vel",
+      "pixels/agentview_image": "observation.images.image",
+      "pixels/robot0_eye_in_hand_image": "observation.images.image2"
+    },
+    "max_parallel_tasks": 1,
+    "disable_env_checker": true,
+    "task_ids": null,
+    "episode_length": null,
+    "obs_type": "pixels_agent_pos",
+    "render_mode": "rgb_array",
+    "camera_name": "agentview_image,robot0_eye_in_hand_image",
+    "init_states": true,
+    "camera_name_mapping": null,
+    "observation_height": 360,
+    "observation_width": 360,
+    "is_libero_plus": false,
+    "control_mode": "relative"
+  },
+  "policy": {
+    "type": "paravla",
+    "n_obs_steps": 1,
+    "input_features": {
+      "observation.images.image": {
+        "type": "VISUAL",
+        "shape": [
+          3,
+          256,
+          256
+        ]
+      },
+      "observation.images.image2": {
+        "type": "VISUAL",
+        "shape": [
+          3,
+          256,
+          256
+        ]
+      },
+      "observation.state": {
+        "type": "STATE",
+        "shape": [
+          8
+        ]
+      }
+    },
+    "output_features": {
+      "action": {
+        "type": "ACTION",
+        "shape": [
+          7
+        ]
+      }
+    },
+    "device": "cuda",
+    "use_amp": true,
+    "use_peft": false,
+    "push_to_hub": false,
+    "repo_id": "GSR/ParaVLA-LIBERO",
+    "private": null,
+    "tags": null,
+    "license": null,
+    "pretrained_path": null,
+    "chunk_size": 16,
+    "n_action_steps": 8,
+    "use_temporal_ensembling": true,
+    "temporal_ensemble_coeff": 0.9,
+    "max_state_dim": 32,
+    "max_action_dim": 32,
+    "normalization_mapping": {
+      "VISUAL": "IDENTITY",
+      "STATE": "MEAN_STD",
+      "ACTION": "MEAN_STD"
+    },
+    "vision_model": "facebook/dinov2-large",
+    "text_model": "google-t5/t5-large",
+    "image_size": 224,
+    "hidden_size": 1024,
+    "fusion_heads": 16,
+    "fusion_layers": 4,
+    "dropout": 0.1,
+    "num_denoise_steps": 20,
+    "num_expert_layers": 8,
+    "num_expert_heads": 16,
+    "expert_head_dim": 64,
+    "expert_injection_stride": 1,
+    "expert_injection_min_scale": 1.0,
+    "use_coord_injection": false,
+    "freeze_vision_encoder": false,
+    "freeze_text_encoder": true,
+    "freeze_patch_embed": false,
+    "unfreeze_blocks": 4,
+    "use_lora": false,
+    "lora_r": 64,
+    "lora_alpha": 128,
+    "lora_dropout": 0.05,
+    "lora_target_modules": [
+      "qkv",
+      "proj",
+      "fc1",
+      "fc2",
+      "patch_embed.proj"
+    ],
+    "tokenizer_max_length": 64,
+    "text_dropout_prob": 0.0,
+    "vision_dim": 1536,
+    "use_goal_predictor": false,
+    "use_wrist_camera": true,
+    "guidance_scale": 1.0,
+    "film_scale_init": 1.0,
+    "optimizer_lr": 0.0001,
+    "backbone_lr_fraction": 0.1,
+    "patch_embed_lr_fraction": 0.5,
+    "optimizer_betas": [
+      0.9,
+      0.95
+    ],
+    "optimizer_eps": 1e-08,
+    "optimizer_weight_decay": 0.0005,
+    "optimizer_grad_clip_norm": 2.0,
+    "scheduler_warmup_steps": 2000,
+    "scheduler_decay_steps": 45000,
+    "scheduler_decay_lr": 1e-06
+  },
+  "reward_model": null,
+  "output_dir": null,
+  "job_name": "paravla_dino_large_wrist",
+  "resume": false,
+  "seed": 1000,
+  "cudnn_deterministic": false,
+  "num_workers": 16,
+  "batch_size": 64,
+  "gradient_accumulation_steps": 1,
+  "prefetch_factor": 8,
+  "persistent_workers": true,
+  "steps": 50000,
+  "eval_freq": -1,
+  "log_freq": 10,
+  "tolerance_s": 0.0001,
+  "save_checkpoint": true,
+  "save_freq": 5000,
+  "use_policy_training_preset": true,
+  "optimizer": {
+    "type": "adamw",
+    "lr": 0.0001,
+    "weight_decay": 0.0005,
+    "grad_clip_norm": 2.0,
+    "betas": [
+      0.9,
+      0.95
+    ],
+    "eps": 1e-08
+  },
+  "scheduler": {
+    "type": "cosine_decay_with_warmup",
+    "num_warmup_steps": 2000,
+    "num_decay_steps": 45000,
+    "peak_lr": 0.0001,
+    "decay_lr": 1e-06
+  },
+  "eval": {
+    "n_episodes": 50,
+    "batch_size": 50,
+    "language_mode": "correct",
+    "language_text": "",
+    "use_async_envs": true
+  },
+  "wandb": {
+    "enable": false,
+    "disable_artifact": false,
+    "project": "lerobot",
+    "entity": null,
+    "notes": null,
+    "run_id": null,
+    "mode": null,
+    "add_tags": true
+  },
+  "peft": null,
+  "sample_weighting": null,
+  "rename_map": {},
+  "checkpoint_path": null
+}