pi05 expert-only fine-tune, batch 32, step 4,800 (pi05_pickblockabs_baseline_bs32_4gpu_20k)

Browse files

Files changed (8) hide show

README.md +60 -0
config.json +103 -0
model.safetensors +3 -0
policy_postprocessor.json +38 -0
policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
policy_preprocessor.json +97 -0
policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
train_config.json +258 -0

README.md ADDED Viewed

	@@ -0,0 +1,60 @@

+---
+library_name: lerobot
+base_model: lerobot/pi05_base
+datasets:
+- Ameyapores/pick_block_eef_position_abs
+tags:
+- robotics
+- pi05
+- lerobot
+- franka
+---
+# π₀.₅ fine-tuned on `pick_block_eef_position_abs` (batch 32, step 4,800)
+[`lerobot/pi05_base`](https://huggingface.co/lerobot/pi05_base) fine-tuned with LeRobot 0.6.1
+(`--policy.type=pi05`) on
+[`Ameyapores/pick_block_eef_position_abs`](https://huggingface.co/datasets/Ameyapores/pick_block_eef_position_abs)
+(35 Franka episodes / 9,181 frames @ 25 fps, 1 task, two 224×224 cameras `cam0`/`cam2`; the
+third π₀.₅ image slot is padded, `empty_cameras=1`; 4-D state).
+This is a **full fine-tune of the action expert** (SigLIP and Gemma-2B frozen,
+`--train_expert_only`, ~0.69B trainable parameters), not a LoRA adapter. The repo holds the
+complete policy weights.
+## Action space — read this before deploying
+Each action is `[x, y, z, gripper]`: the **absolute end-effector position to reach at the next
+frame** (same frame and units as `observation.state[:3]`) plus a binary gripper target (0/1).
+It is not interchangeable with delta-action models such as
+[`arkojit1/pi05_pick_block_eef_delta`](https://huggingface.co/arkojit1/pi05_pick_block_eef_delta).
+The dataset is byte-identical to the local build behind
+[`arkojit1/pi05_pick_block_eef_abs`](https://huggingface.co/arkojit1/pi05_pick_block_eef_abs).
+State and action use quantile (q01–q99) normalisation. The training region is narrow:
+x 0.528–0.559, y 0.056–0.068, z 0.145–0.311, so targets outside it are extrapolation.
+## Training
+| setting | value |
+|---|---|
+| trainable | action expert + projections (~0.69B); SigLIP and Gemma-2B frozen |
+| batch | 32 global (4 GPUs × 8) |
+| optimizer | AdamW, peak LR 2.5e-5, cosine to 2.5e-6, schedule pinned to 20,000 steps, 666 warmup |
+| this checkpoint | step **4,800** (~19 epochs) of a 20,000-step run |
+| other | bf16, gradient checkpointing, `torch.compile` (default mode), image augmentation on, chunk_size 50, n_action_steps 50 |
+## Evaluation
+Flow-matching loss on 4 held-out episodes (31–34), **not** a task success rate:
+- this checkpoint: eval **0.0576**, train 0.056. This is the best eval of the run and the
+  lowest of any run on this dataset.
+- the run continued to 20,000 steps (79 epochs) and overfit: eval 0.0839, train 0.034.
+- point-to-point eval scatter on this 4-episode holdout is about ±0.005–0.01.
+## Loading
+```bash
+lerobot-eval --policy.path=arkojit1/pi05_pick_block_eef_abs_bs32_lr2.5e-5_4.8k ...
+```

config.json ADDED Viewed

	@@ -0,0 +1,103 @@

+{
+    "type": "pi05",
+    "n_obs_steps": 1,
+    "input_features": {
+        "observation.images.cam0": {
+            "type": "VISUAL",
+            "shape": [
+                3,
+                224,
+                224
+            ]
+        },
+        "observation.images.cam2": {
+            "type": "VISUAL",
+            "shape": [
+                3,
+                224,
+                224
+            ]
+        },
+        "observation.state": {
+            "type": "STATE",
+            "shape": [
+                4
+            ]
+        },
+        "observation.images.empty_camera_0": {
+            "type": "VISUAL",
+            "shape": [
+                3,
+                224,
+                224
+            ]
+        }
+    },
+    "output_features": {
+        "action": {
+            "type": "ACTION",
+            "shape": [
+                4
+            ]
+        }
+    },
+    "device": "cuda",
+    "use_amp": false,
+    "use_peft": false,
+    "push_to_hub": false,
+    "repo_id": null,
+    "private": null,
+    "tags": null,
+    "license": null,
+    "pretrained_path": "runs/pi05_pickblockabs_baseline_bs32_4gpu_20k-20261006-191836/checkpoints/last/pretrained_model",
+    "pretrained_revision": null,
+    "paligemma_variant": "gemma_2b",
+    "action_expert_variant": "gemma_300m",
+    "dtype": "bfloat16",
+    "chunk_size": 50,
+    "n_action_steps": 50,
+    "max_state_dim": 32,
+    "max_action_dim": 32,
+    "num_inference_steps": 10,
+    "time_sampling_beta_alpha": 1.5,
+    "time_sampling_beta_beta": 1.0,
+    "time_sampling_scale": 0.999,
+    "time_sampling_offset": 0.001,
+    "min_period": 0.004,
+    "max_period": 4.0,
+    "use_relative_actions": false,
+    "relative_exclude_joints": [
+        "gripper"
+    ],
+    "action_feature_names": [
+        "axes"
+    ],
+    "rtc_config": null,
+    "image_resolution": [
+        224,
+        224
+    ],
+    "empty_cameras": 1,
+    "tokenizer_max_length": 200,
+    "normalization_mapping": {
+        "VISUAL": "IDENTITY",
+        "STATE": "QUANTILES",
+        "ACTION": "QUANTILES"
+    },
+    "gradient_checkpointing": true,
+    "compile_model": true,
+    "compile_mode": "default",
+    "freeze_vision_encoder": false,
+    "train_expert_only": true,
+    "optimizer_lr": 2.5e-05,
+    "optimizer_betas": [
+        0.9,
+        0.95
+    ],
+    "optimizer_eps": 1e-08,
+    "optimizer_weight_decay": 0.01,
+    "optimizer_grad_clip_norm": 1.0,
+    "scheduler_warmup_steps": 666,
+    "scheduler_decay_steps": 20000,
+    "scheduler_decay_lr": 2.5e-06
+}

model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:792a0d278b9323787757a3266766c1aecb61120b7c905ecf321397b1eb25a2b6
+size 9354050752

policy_postprocessor.json ADDED Viewed

	@@ -0,0 +1,38 @@

+{
+  "name": "policy_postprocessor",
+  "steps": [
+    {
+      "registry_name": "unnormalizer_processor",
+      "config": {
+        "eps": 1e-08,
+        "features": {
+          "action": {
+            "type": "ACTION",
+            "shape": [
+              4
+            ]
+          }
+        },
+        "norm_map": {
+          "VISUAL": "IDENTITY",
+          "STATE": "QUANTILES",
+          "ACTION": "QUANTILES"
+        }
+      },
+      "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
+    },
+    {
+      "registry_name": "absolute_actions_processor",
+      "config": {
+        "enabled": false
+      }
+    },
+    {
+      "registry_name": "device_processor",
+      "config": {
+        "device": "cpu",
+        "float_dtype": null
+      }
+    }
+  ]
+}

policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:4e44ccb252f706a61f8702be2474276b532a28010f88f50148cecdde906ac2ae
+size 7384

policy_preprocessor.json ADDED Viewed

	@@ -0,0 +1,97 @@

+{
+  "name": "policy_preprocessor",
+  "steps": [
+    {
+      "registry_name": "rename_observations_processor",
+      "config": {
+        "rename_map": {}
+      }
+    },
+    {
+      "registry_name": "to_batch_processor",
+      "config": {}
+    },
+    {
+      "registry_name": "relative_actions_processor",
+      "config": {
+        "enabled": false,
+        "exclude_joints": [
+          "gripper"
+        ],
+        "action_names": null
+      }
+    },
+    {
+      "registry_name": "normalizer_processor",
+      "config": {
+        "eps": 1e-08,
+        "features": {
+          "observation.images.cam0": {
+            "type": "VISUAL",
+            "shape": [
+              3,
+              224,
+              224
+            ]
+          },
+          "observation.images.cam2": {
+            "type": "VISUAL",
+            "shape": [
+              3,
+              224,
+              224
+            ]
+          },
+          "observation.state": {
+            "type": "STATE",
+            "shape": [
+              4
+            ]
+          },
+          "observation.images.empty_camera_0": {
+            "type": "VISUAL",
+            "shape": [
+              3,
+              224,
+              224
+            ]
+          },
+          "action": {
+            "type": "ACTION",
+            "shape": [
+              4
+            ]
+          }
+        },
+        "norm_map": {
+          "VISUAL": "IDENTITY",
+          "STATE": "QUANTILES",
+          "ACTION": "QUANTILES"
+        }
+      },
+      "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
+    },
+    {
+      "registry_name": "pi05_prepare_state_tokenizer_processor_step",
+      "config": {}
+    },
+    {
+      "registry_name": "tokenizer_processor",
+      "config": {
+        "max_length": 200,
+        "task_key": "task",
+        "padding_side": "right",
+        "padding": "max_length",
+        "truncation": true,
+        "tokenizer_name": "google/paligemma-3b-pt-224"
+      }
+    },
+    {
+      "registry_name": "device_processor",
+      "config": {
+        "device": "cuda",
+        "float_dtype": null
+      }
+    }
+  ]
+}

policy_preprocessor_step_3_normalizer_processor.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:42453420ca049c1de5634a98f4c01a833325fa1c6a0ae87ffedd96e2dfbd09db
+size 7328

train_config.json ADDED Viewed

	@@ -0,0 +1,258 @@

+{
+    "dataset": {
+        "repo_id": "Ameyapores/pick_block_eef_position_abs",
+        "repo_type": "dataset",
+        "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": "pyav",
+        "return_uint8": false,
+        "depth_output_unit": "mm",
+        "streaming": false,
+        "eval_split": 0.1
+    },
+    "env": null,
+    "policy": {
+        "type": "pi05",
+        "n_obs_steps": 1,
+        "input_features": {
+            "observation.images.cam0": {
+                "type": "VISUAL",
+                "shape": [
+                    3,
+                    224,
+                    224
+                ]
+            },
+            "observation.images.cam2": {
+                "type": "VISUAL",
+                "shape": [
+                    3,
+                    224,
+                    224
+                ]
+            },
+            "observation.state": {
+                "type": "STATE",
+                "shape": [
+                    4
+                ]
+            },
+            "observation.images.empty_camera_0": {
+                "type": "VISUAL",
+                "shape": [
+                    3,
+                    224,
+                    224
+                ]
+            }
+        },
+        "output_features": {
+            "action": {
+                "type": "ACTION",
+                "shape": [
+                    4
+                ]
+            }
+        },
+        "device": "cuda",
+        "use_amp": false,
+        "use_peft": false,
+        "push_to_hub": false,
+        "repo_id": null,
+        "private": null,
+        "tags": null,
+        "license": null,
+        "pretrained_path": "runs/pi05_pickblockabs_baseline_bs32_4gpu_20k-20261006-191836/checkpoints/last/pretrained_model",
+        "pretrained_revision": null,
+        "paligemma_variant": "gemma_2b",
+        "action_expert_variant": "gemma_300m",
+        "dtype": "bfloat16",
+        "chunk_size": 50,
+        "n_action_steps": 50,
+        "max_state_dim": 32,
+        "max_action_dim": 32,
+        "num_inference_steps": 10,
+        "time_sampling_beta_alpha": 1.5,
+        "time_sampling_beta_beta": 1.0,
+        "time_sampling_scale": 0.999,
+        "time_sampling_offset": 0.001,
+        "min_period": 0.004,
+        "max_period": 4.0,
+        "use_relative_actions": false,
+        "relative_exclude_joints": [
+            "gripper"
+        ],
+        "action_feature_names": [
+            "axes"
+        ],
+        "rtc_config": null,
+        "image_resolution": [
+            224,
+            224
+        ],
+        "empty_cameras": 1,
+        "tokenizer_max_length": 200,
+        "normalization_mapping": {
+            "VISUAL": "IDENTITY",
+            "STATE": "QUANTILES",
+            "ACTION": "QUANTILES"
+        },
+        "gradient_checkpointing": true,
+        "compile_model": true,
+        "compile_mode": "default",
+        "freeze_vision_encoder": false,
+        "train_expert_only": true,
+        "optimizer_lr": 2.5e-05,
+        "optimizer_betas": [
+            0.9,
+            0.95
+        ],
+        "optimizer_eps": 1e-08,
+        "optimizer_weight_decay": 0.01,
+        "optimizer_grad_clip_norm": 1.0,
+        "scheduler_warmup_steps": 666,
+        "scheduler_decay_steps": 20000,
+        "scheduler_decay_lr": 2.5e-06
+    },
+    "reward_model": null,
+    "output_dir": "runs/pi05_pickblockabs_baseline_bs32_4gpu_20k-20261006-191836",
+    "job_name": "pi05_pickblockabs_baseline_bs32_4gpu_20k-20261006-191836",
+    "resume": true,
+    "seed": 1000,
+    "cudnn_deterministic": false,
+    "num_workers": 2,
+    "batch_size": 8,
+    "prefetch_factor": 4,
+    "persistent_workers": true,
+    "dataloader_multiprocessing_context": "spawn",
+    "steps": 20000,
+    "env_eval_freq": 20000,
+    "log_freq": 200,
+    "eval_steps": 200,
+    "max_eval_samples": 0,
+    "tolerance_s": 0.0001,
+    "save_checkpoint": true,
+    "save_freq": 200,
+    "use_policy_training_preset": true,
+    "optimizer": {
+        "type": "adamw",
+        "lr": 2.5e-05,
+        "weight_decay": 0.01,
+        "grad_clip_norm": 1.0,
+        "betas": [
+            0.9,
+            0.95
+        ],
+        "eps": 1e-08
+    },
+    "scheduler": {
+        "type": "cosine_decay_with_warmup",
+        "num_warmup_steps": 666,
+        "num_decay_steps": 20000,
+        "peak_lr": 2.5e-05,
+        "decay_lr": 2.5e-06
+    },
+    "eval": {
+        "n_episodes": 50,
+        "batch_size": 50,
+        "use_async_envs": true,
+        "recording": false,
+        "recording_repo_id": null,
+        "recording_private": false
+    },
+    "wandb": {
+        "enable": false,
+        "disable_artifact": false,
+        "project": "lerobot",
+        "entity": null,
+        "notes": null,
+        "run_id": null,
+        "mode": null,
+        "add_tags": true
+    },
+    "peft": null,
+    "job": {
+        "target": null,
+        "image": "huggingface/lerobot-gpu:latest",
+        "timeout": "2d",
+        "detach": false,
+        "tags": []
+    },
+    "save_checkpoint_to_hub": false,
+    "sample_weighting": null,
+    "rename_map": {},
+    "checkpoint_path": "runs/pi05_pickblockabs_baseline_bs32_4gpu_20k-20261006-191836/checkpoints/last"
+}