chore(ci): make quality checks pass across repo

- fix ruff violations in dataset, recording, and value modules

- fix typos findings and markdown/workflow lint issues

- update pages workflow checkout to disable credential persistence

- apply pre-commit auto-formatting and EOF normalization

- include requested repo housekeeping updates (.gitignore and README)
这个提交包含在:
Elvin
2026-03-04 11:02:06 +08:00
父节点 8b6a85cf0f
当前提交 c96c678772
共有 26 个文件被更改,包括 143 次插入 和 115 次删除

查看文件

@ -29,6 +29,8 @@ jobs:
steps: steps:
- name: Checkout - name: Checkout
uses: actions/checkout@v4 uses: actions/checkout@v4
with:
persist-credentials: false
- name: Setup Pages - name: Setup Pages
uses: actions/configure-pages@v5 uses: actions/configure-pages@v5

1
.gitignore vendored
查看文件

@ -177,3 +177,4 @@ outputs/
*.urdf *.urdf
*.xml *.xml
*.part *.part
AGENTS.md

查看文件

@ -22,7 +22,7 @@ If using `conda activate /llm_jzm/cache/conda_env/lerobot`:
- Treat it as shared machine mode. - Treat it as shared machine mode.
- Hugging Face cache is not in the default location. - Hugging Face cache is not in the default location.
- Set Hugging Face token to the shared token for runs on this machine. - Set Hugging Face token to the shared token for runs on this machine.
- Set Wandb token to hte shared token for runs on this machine - Set Wandb token to the shared token for runs on this machine
Use: Use:

查看文件

@ -1,7 +1,7 @@
<h1 align="center">Evo-RL</h1> <h1 align="center">Evo-RL</h1>
<p align="center"> <p align="center">
<a href="https://elvin-yk.github.io/Evo-RL/"><img alt="project website" src="https://img.shields.io/badge/Project-Website-0ea5e9"/></a> <a href="https://MINT-SJTU.github.io/Evo-RL/"><img alt="project website" src="https://img.shields.io/badge/Project-Website-0ea5e9"/></a>
<a href="https://github.com/huggingface/lerobot"><img alt="lerobot version" src="https://img.shields.io/badge/LeRobot-0.4.4-f59e0b"/></a> <a href="https://github.com/huggingface/lerobot"><img alt="lerobot version" src="https://img.shields.io/badge/LeRobot-0.4.4-f59e0b"/></a>
<a href="https://evorl.example.com/wechat-post"><img alt="wechat post" src="https://img.shields.io/badge/WeChat-Official%20Post-07c160"/></a> <a href="https://evorl.example.com/wechat-post"><img alt="wechat post" src="https://img.shields.io/badge/WeChat-Official%20Post-07c160"/></a>
<a href="#community-channels"><img alt="wechat group join us" src="https://img.shields.io/badge/WeChat%20Group-Join%20Us-a855f7?logo=wechat&logoColor=white"/></a> <a href="#community-channels"><img alt="wechat group join us" src="https://img.shields.io/badge/WeChat%20Group-Join%20Us-a855f7?logo=wechat&logoColor=white"/></a>
@ -84,6 +84,7 @@ For SO101 setup, please follow the [official SO101 tutorial](https://wiki.seeeds
#### Device path recommendation #### Device path recommendation
Recommended path strategy: Recommended path strategy:
- **Robot serial:** use `/dev/serial/by-id/` (stable across reboots). - **Robot serial:** use `/dev/serial/by-id/` (stable across reboots).
- **Cameras:** prefer `/dev/v4l/by-id/`; if IDs are not unique, use `/dev/v4l/by-path/`. - **Cameras:** prefer `/dev/v4l/by-id/`; if IDs are not unique, use `/dev/v4l/by-path/`.
- In examples below: robot ports use `by-id`, camera paths use `by-path`. - In examples below: robot ports use `by-id`, camera paths use `by-path`.
@ -213,6 +214,7 @@ lerobot-human-inloop-record \
Recommendation: use **`fourcc: "MJPG"`** for OpenCV and **`warmup_s`** for RealSense. In this example `front` uses RealSense, but you can switch it to OpenCV with the same structure. Recommendation: use **`fourcc: "MJPG"`** for OpenCV and **`warmup_s`** for RealSense. In this example `front` uses RealSense, but you can switch it to OpenCV with the same structure.
Hotkeys: Hotkeys:
- `i`: toggle intervention mode (policy <-> teleop takeover) - `i`: toggle intervention mode (policy <-> teleop takeover)
- `s`: mark success and end current episode - `s`: mark success and end current episode
- `f`: mark failure and end current episode - `f`: mark failure and end current episode
@ -230,7 +232,7 @@ This prints: dataset meta, totals, episode-length stats/histogram, success/inter
### 4) Value Function Training ### 4) Value Function Training
Train the value function on the current dataset. Current default: [Pi*0.6](https://www.pi.website/blog/pistar06) (`--value.type=pistar06`). Train the value function on the current dataset. Current default: [Pi\*0.6](https://www.pi.website/blog/pistar06) (`--value.type=pistar06`).
Single-GPU template: Single-GPU template:
@ -260,6 +262,7 @@ CUDA_VISIBLE_DEVICES=<GPU_ID_LIST> accelerate launch \
``` ```
To plug in a different value function, minimal path in this repo: To plug in a different value function, minimal path in this repo:
- Add `src/lerobot/values/<your_value>/configuration_<your_value>.py` with `@PreTrainedConfig.register_subclass("<your_value>")`. - Add `src/lerobot/values/<your_value>/configuration_<your_value>.py` with `@PreTrainedConfig.register_subclass("<your_value>")`.
- Add `src/lerobot/values/<your_value>/modeling_<your_value>.py` with `<YourValue>Policy(PreTrainedPolicy)` (implement at least `forward`, `predict_value`, and `build_training_raw_batch_hook` for `lerobot-value-train`). - Add `src/lerobot/values/<your_value>/modeling_<your_value>.py` with `<YourValue>Policy(PreTrainedPolicy)` (implement at least `forward`, `predict_value`, and `build_training_raw_batch_hook` for `lerobot-value-train`).
- Add `src/lerobot/values/<your_value>/processor_<your_value>.py` with `make_<your_value>_pre_post_processors(...)`. - Add `src/lerobot/values/<your_value>/processor_<your_value>.py` with `make_<your_value>_pre_post_processors(...)`.
@ -268,6 +271,7 @@ To plug in a different value function, minimal path in this repo:
### 5) Value Inference ### 5) Value Inference
Infer value signals and write value/advantage/indicator back to the dataset: Infer value signals and write value/advantage/indicator back to the dataset:
- `value`: estimated return-to-go of the current frame. - `value`: estimated return-to-go of the current frame.
- `advantage`: relative improvement signal (higher means better-than-baseline trajectory quality). - `advantage`: relative improvement signal (higher means better-than-baseline trajectory quality).
- `indicator`: binarized training tag derived from advantage. - `indicator`: binarized training tag derived from advantage.
@ -302,10 +306,12 @@ CUDA_VISIBLE_DEVICES=<GPU_ID_LIST> accelerate launch \
``` ```
Parameter notes: Parameter notes:
- `--acp.n_step`: n-step advantage horizon. - `--acp.n_step`: n-step advantage horizon.
- `--acp.positive_ratio`: positive label ratio after advantage binarization (e.g., `0.3` = top 30% per task). - `--acp.positive_ratio`: positive label ratio after advantage binarization (e.g., `0.3` = top 30% per task).
Expected new columns: Expected new columns:
- `complementary_info.value_<TAG>` - `complementary_info.value_<TAG>`
- `complementary_info.advantage_<TAG>` - `complementary_info.advantage_<TAG>`
- `complementary_info.acp_indicator_<TAG>` - `complementary_info.acp_indicator_<TAG>`
@ -313,6 +319,7 @@ Expected new columns:
These columns are written back to the original dataset specified by `--dataset.repo_id`. These columns are written back to the original dataset specified by `--dataset.repo_id`.
Visualization (optional): Visualization (optional):
- Enable with `--viz.enable=true`. - Enable with `--viz.enable=true`.
- Output videos are saved to `outputs/value_infer/<RUN_NAME>/value/viz/`. - Output videos are saved to `outputs/value_infer/<RUN_NAME>/value/viz/`.
- Each video overlays a value curve and per-frame text (`advantage`, `acp_indicator`) on top of camera frames. - Each video overlays a value curve and per-frame text (`advantage`, `acp_indicator`) on top of camera frames.
@ -323,6 +330,7 @@ Visualization (optional):
- `--viz.vcodec=<CODEC>` and `--viz.frame_storage_mode=memory|disk` - `--viz.vcodec=<CODEC>` and `--viz.frame_storage_mode=memory|disk`
Value overlay examples (SO101, smooth=11): Value overlay examples (SO101, smooth=11):
- GIF assets are stored under `website/assets/gifs/`. - GIF assets are stored under `website/assets/gifs/`.
### 6) Policy Training ### 6) Policy Training
@ -354,6 +362,7 @@ lerobot-train \
`--acp.indicator_dropout_prob` controls tag drop rate in task text; `0.3` helps learn both tagged and untagged conditions. `--acp.indicator_dropout_prob` controls tag drop rate in task text; `0.3` helps learn both tagged and untagged conditions.
Important checks: Important checks:
- `--acp.indicator_field` must exist in the dataset and be **binary (`0/1`)**. - `--acp.indicator_field` must exist in the dataset and be **binary (`0/1`)**.
Multi-GPU template: Multi-GPU template:
@ -396,6 +405,7 @@ lerobot-human-inloop-record \
``` ```
Dataset continuation options: Dataset continuation options:
- **Append in place:** keep `--resume=true` and continue recording into the same dataset. - **Append in place:** keep `--resume=true` and continue recording into the same dataset.
- **Merge multiple rounds:** use the official dataset editor to merge separate datasets. - **Merge multiple rounds:** use the official dataset editor to merge separate datasets.
@ -407,6 +417,7 @@ lerobot-edit-dataset \
``` ```
Additional data attributes vs default `lerobot-record` behavior: Additional data attributes vs default `lerobot-record` behavior:
- `complementary_info.policy_action`: policy output action at each step. - `complementary_info.policy_action`: policy output action at each step.
- `complementary_info.is_intervention`: whether current step is in intervention. - `complementary_info.is_intervention`: whether current step is in intervention.
- `complementary_info.state`: intervention state-machine state. - `complementary_info.state`: intervention state-machine state.

查看文件

@ -41,7 +41,9 @@ class ValueTargetsConfig:
if self.c_fail_coef < 0: if self.c_fail_coef < 0:
raise ValueError("'targets.c_fail_coef' must be non-negative.") raise ValueError("'targets.c_fail_coef' must be non-negative.")
if not self.target_field.startswith("observation."): if not self.target_field.startswith("observation."):
raise ValueError("'targets.target_field' must start with 'observation.' to survive processor conversion.") raise ValueError(
"'targets.target_field' must start with 'observation.' to survive processor conversion."
)
@dataclass @dataclass
@ -113,7 +115,7 @@ class ValueTrainPipelineConfig(HubMixin):
self.targets.validate() self.targets.validate()
if hasattr(self.value, "target_key"): if hasattr(self.value, "target_key"):
setattr(self.value, "target_key", self.targets.target_field) self.value.target_key = self.targets.target_field
if not self.job_name: if not self.job_name:
self.job_name = f"{self.value.type}_value" self.job_name = f"{self.value.type}_value"
@ -136,7 +138,9 @@ class ValueTrainPipelineConfig(HubMixin):
self.scheduler = self.value.get_scheduler_preset() self.scheduler = self.value.get_scheduler_preset()
if self.value.push_to_hub and not self.value.repo_id: if self.value.push_to_hub and not self.value.repo_id:
raise ValueError("'value.repo_id' argument missing. Please specify it to push the model to the hub.") raise ValueError(
"'value.repo_id' argument missing. Please specify it to push the model to the hub."
)
@classmethod @classmethod
def __get_path_fields__(cls) -> list[str]: def __get_path_fields__(cls) -> list[str]:

查看文件

@ -836,7 +836,11 @@ def _copy_and_reindex_episodes_metadata(
for key in schema.names: for key in schema.names:
if key in {"episode_index", "tasks", "length"}: if key in {"episode_index", "tasks", "length"}:
continue continue
if key.startswith("stats/") or key.startswith("__index_level_") or key.startswith("meta/episodes/"): if (
key.startswith("stats/")
or key.startswith("__index_level_")
or key.startswith("meta/episodes/")
):
continue continue
if key not in seen_preserved_keys: if key not in seen_preserved_keys:
seen_preserved_keys.add(key) seen_preserved_keys.add(key)

查看文件

@ -729,15 +729,14 @@ class LeRobotDataset(torch.utils.data.Dataset):
# Check if cached dataset contains all requested episodes # Check if cached dataset contains all requested episodes
if not self._check_cached_episodes_sufficient(): if not self._check_cached_episodes_sufficient():
raise FileNotFoundError("Cached dataset doesn't contain all requested episodes") raise FileNotFoundError("Cached dataset doesn't contain all requested episodes")
except (AssertionError, FileNotFoundError, NotADirectoryError): except (AssertionError, FileNotFoundError, NotADirectoryError) as err:
if is_valid_version(self.revision): if is_valid_version(self.revision):
self.revision = get_safe_version(self.repo_id, self.revision) self.revision = get_safe_version(self.repo_id, self.revision)
self.download(download_videos) self.download(download_videos)
self.hf_dataset = self.load_hf_dataset() self.hf_dataset = self.load_hf_dataset()
if not self._check_cached_episodes_sufficient(): if not self._check_cached_episodes_sufficient():
missing_videos_hint = ( missing_videos_hint = (
" Missing video files are required by this dataset. " " Missing video files are required by this dataset. Retry with `download_videos=True`."
"Retry with `download_videos=True`."
if len(self.meta.video_keys) > 0 and not download_videos if len(self.meta.video_keys) > 0 and not download_videos
else "" else ""
) )
@ -745,7 +744,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
"Cached dataset is incomplete after download. " "Cached dataset is incomplete after download. "
"Some requested episodes or files are missing." "Some requested episodes or files are missing."
f"{missing_videos_hint}" f"{missing_videos_hint}"
) ) from err
# Create mapping from absolute indices to relative indices when only a subset of the episodes are loaded # Create mapping from absolute indices to relative indices when only a subset of the episodes are loaded
# Build a mapping: absolute_index -> relative_index_in_filtered_dataset # Build a mapping: absolute_index -> relative_index_in_filtered_dataset

查看文件

@ -60,10 +60,7 @@ def _from_meta_stats(dataset: Any, indicator_field: str) -> ACPIndicatorStats |
count_raw = _stats_entry_to_scalar(field_stats, "count") count_raw = _stats_entry_to_scalar(field_stats, "count")
total_count = int(round(count_raw)) if count_raw is not None else -1 total_count = int(round(count_raw)) if count_raw is not None else -1
if total_count >= 0: positive_count = int(round(ratio * total_count)) if total_count >= 0 else -1
positive_count = int(round(ratio * total_count))
else:
positive_count = -1
return ACPIndicatorStats( return ACPIndicatorStats(
indicator_field=indicator_field, indicator_field=indicator_field,

查看文件

@ -136,7 +136,9 @@ def build_report(dataset_root: Path) -> dict[str, Any]:
episodes_df = episodes_ds.to_pandas() episodes_df = episodes_ds.to_pandas()
actual_episode_count = int(len(episodes_df)) actual_episode_count = int(len(episodes_df))
episode_success_labels = episodes_df["episode_success"].tolist() if "episode_success" in episodes_df else [] episode_success_labels = (
episodes_df["episode_success"].tolist() if "episode_success" in episodes_df else []
)
normalized_labels = [_normalize_label(v) for v in episode_success_labels] normalized_labels = [_normalize_label(v) for v in episode_success_labels]
success_count = sum(1 for v in normalized_labels if v == "success") success_count = sum(1 for v in normalized_labels if v == "success")
@ -153,7 +155,11 @@ def build_report(dataset_root: Path) -> dict[str, Any]:
intervention_frames = 0 intervention_frames = 0
intervention_episode_count = 0 intervention_episode_count = 0
if intervention_col in data_schema_names and episode_index_col in data_schema_names and actual_frame_count > 0: if (
intervention_col in data_schema_names
and episode_index_col in data_schema_names
and actual_frame_count > 0
):
table = data_dataset.to_table(columns=[episode_index_col, intervention_col]) table = data_dataset.to_table(columns=[episode_index_col, intervention_col])
episode_indices = table[episode_index_col].to_pylist() episode_indices = table[episode_index_col].to_pylist()
intervention_values = table[intervention_col].to_pylist() intervention_values = table[intervention_col].to_pylist()
@ -190,10 +196,14 @@ def build_report(dataset_root: Path) -> dict[str, Any]:
fps = float(info.get("fps", 0) or 0) fps = float(info.get("fps", 0) or 0)
episode_lengths_frames = ( episode_lengths_frames = (
episodes_df["length"].astype(int).tolist() if "length" in episodes_df and actual_episode_count > 0 else [] episodes_df["length"].astype(int).tolist()
if "length" in episodes_df and actual_episode_count > 0
else []
) )
episode_lengths_seconds = [frames / fps for frames in episode_lengths_frames] if fps > 0 else [] episode_lengths_seconds = [frames / fps for frames in episode_lengths_frames] if fps > 0 else []
length_mean = float(sum(episode_lengths_seconds) / len(episode_lengths_seconds)) if episode_lengths_seconds else 0.0 length_mean = (
float(sum(episode_lengths_seconds) / len(episode_lengths_seconds)) if episode_lengths_seconds else 0.0
)
length_min = float(min(episode_lengths_seconds)) if episode_lengths_seconds else 0.0 length_min = float(min(episode_lengths_seconds)) if episode_lengths_seconds else 0.0
length_max = float(max(episode_lengths_seconds)) if episode_lengths_seconds else 0.0 length_max = float(max(episode_lengths_seconds)) if episode_lengths_seconds else 0.0
length_histogram = _build_episode_length_histogram(episode_lengths_seconds, bins=20) length_histogram = _build_episode_length_histogram(episode_lengths_seconds, bins=20)

查看文件

@ -22,16 +22,16 @@ Phase A/B goals:
- Keyboard-toggled intervention lets teleop temporarily take over execution. - Keyboard-toggled intervention lets teleop temporarily take over execution.
""" """
import logging
import json import json
import logging
import time import time
from pathlib import Path from pathlib import Path
from typing import Any from typing import Any
from lerobot.configs import parser from lerobot.configs import parser
from lerobot.scripts.lerobot_record import RecordConfig, record from lerobot.scripts.lerobot_record import RecordConfig, record
from lerobot.utils.import_utils import register_third_party_plugins
from lerobot.utils.constants import HF_LEROBOT_HOME from lerobot.utils.constants import HF_LEROBOT_HOME
from lerobot.utils.import_utils import register_third_party_plugins
from lerobot.utils.recording_annotations import EPISODE_FAILURE, infer_collector_policy_version from lerobot.utils.recording_annotations import EPISODE_FAILURE, infer_collector_policy_version
@ -64,7 +64,9 @@ def _load_failure_reset_pose(pose_path: Path) -> dict[str, float]:
payload = json.load(f) payload = json.load(f)
joint_pos_raw = payload["joint_pos"] if isinstance(payload, dict) and "joint_pos" in payload else payload joint_pos_raw = payload["joint_pos"] if isinstance(payload, dict) and "joint_pos" in payload else payload
if not isinstance(joint_pos_raw, dict): if not isinstance(joint_pos_raw, dict):
raise ValueError(f"Invalid failure reset pose payload in {pose_path}: expected dict, got {type(joint_pos_raw)}") raise ValueError(
f"Invalid failure reset pose payload in {pose_path}: expected dict, got {type(joint_pos_raw)}"
)
joint_pos = {str(key): float(value) for key, value in joint_pos_raw.items() if str(key).endswith(".pos")} joint_pos = {str(key): float(value) for key, value in joint_pos_raw.items() if str(key).endswith(".pos")}
if not joint_pos: if not joint_pos:
raise ValueError(f"Invalid failure reset pose payload in {pose_path}: no '.pos' joints found.") raise ValueError(f"Invalid failure reset pose payload in {pose_path}: no '.pos' joints found.")

查看文件

@ -97,6 +97,13 @@ from lerobot.robots import ( # noqa: F401
so_follower, so_follower,
unitree_g1 as unitree_g1_robot, unitree_g1 as unitree_g1_robot,
) )
from lerobot.scripts.recording_hil import (
ACPInferenceConfig,
PolicySyncDualArmExecutor,
_capture_policy_runtime_state, # noqa: F401
_predict_policy_action_with_acp_inference, # noqa: F401
)
from lerobot.scripts.recording_loop import record_loop
from lerobot.teleoperators import ( # noqa: F401 from lerobot.teleoperators import ( # noqa: F401
TeleoperatorConfig, TeleoperatorConfig,
bi_openarm_leader, bi_openarm_leader,
@ -110,13 +117,6 @@ from lerobot.teleoperators import ( # noqa: F401
so_leader, so_leader,
unitree_g1, unitree_g1,
) )
from lerobot.scripts.recording_hil import (
ACPInferenceConfig,
PolicySyncDualArmExecutor,
_capture_policy_runtime_state, # noqa: F401
_predict_policy_action_with_acp_inference, # noqa: F401
)
from lerobot.scripts.recording_loop import record_loop
from lerobot.utils.constants import ACTION from lerobot.utils.constants import ACTION
from lerobot.utils.control_utils import ( from lerobot.utils.control_utils import (
init_keyboard_listener, init_keyboard_listener,
@ -288,6 +288,7 @@ class RecordConfig:
"""This enables the parser to load config from the policy using `--policy.path=local/dir`""" """This enables the parser to load config from the policy using `--policy.path=local/dir`"""
return ["policy"] return ["policy"]
@parser.wrap() @parser.wrap()
def record(cfg: RecordConfig) -> LeRobotDataset: def record(cfg: RecordConfig) -> LeRobotDataset:
init_logging() init_logging()
@ -325,10 +326,7 @@ def record(cfg: RecordConfig) -> LeRobotDataset:
) )
if cfg.intervention_state_machine_enabled and cfg.policy is not None and cfg.teleop is not None: if cfg.intervention_state_machine_enabled and cfg.policy is not None and cfg.teleop is not None:
action_names = dataset_features[ACTION]["names"] action_names = dataset_features[ACTION]["names"]
if action_names is None: action_names = list(robot.action_features) if action_names is None else list(action_names)
action_names = list(robot.action_features)
else:
action_names = list(action_names)
dataset_features["complementary_info.policy_action"] = { dataset_features["complementary_info.policy_action"] = {
"dtype": "float32", "dtype": "float32",
"shape": (len(action_names),), "shape": (len(action_names),),
@ -524,9 +522,7 @@ def record(cfg: RecordConfig) -> LeRobotDataset:
continue continue
extra_episode_metadata = ( extra_episode_metadata = (
{"episode_success": episode_success} {"episode_success": episode_success} if cfg.enable_episode_outcome_labeling else None
if cfg.enable_episode_outcome_labeling
else None
) )
dataset.save_episode(extra_episode_metadata=extra_episode_metadata) dataset.save_episode(extra_episode_metadata=extra_episode_metadata)
recorded_episodes += 1 recorded_episodes += 1

查看文件

@ -35,6 +35,9 @@ from lerobot.configs.value import ValueInferencePipelineConfig
from lerobot.datasets.lerobot_dataset import LeRobotDataset from lerobot.datasets.lerobot_dataset import LeRobotDataset
from lerobot.datasets.utils import load_info, write_info from lerobot.datasets.utils import load_info, write_info
from lerobot.policies.factory import make_policy, make_pre_post_processors from lerobot.policies.factory import make_policy, make_pre_post_processors
from lerobot.scripts.value_infer_viz import (
_export_overlay_videos,
)
from lerobot.utils.constants import ( from lerobot.utils.constants import (
CHECKPOINTS_DIR, CHECKPOINTS_DIR,
LAST_CHECKPOINT_LINK, LAST_CHECKPOINT_LINK,
@ -44,11 +47,6 @@ from lerobot.utils.import_utils import register_third_party_plugins
from lerobot.utils.random_utils import set_seed from lerobot.utils.random_utils import set_seed
from lerobot.utils.recording_annotations import EPISODE_SUCCESS, resolve_episode_success_label from lerobot.utils.recording_annotations import EPISODE_SUCCESS, resolve_episode_success_label
from lerobot.utils.utils import init_logging, inside_slurm from lerobot.utils.utils import init_logging, inside_slurm
from lerobot.scripts.value_infer_viz import (
_export_overlay_videos,
_get_episode_value_bounds,
_get_video_encode_options,
)
from lerobot.values.pistar06.configuration_pistar06 import Pistar06Config from lerobot.values.pistar06.configuration_pistar06 import Pistar06Config
from lerobot.values.pistar06.modeling_pistar06 import ( from lerobot.values.pistar06.modeling_pistar06 import (
EpisodeTargetInfo, EpisodeTargetInfo,
@ -310,8 +308,8 @@ def _binarize_advantages(
for i in range(advantages.shape[0]): for i in range(advantages.shape[0]):
task_idx = int(task_indices[i]) task_idx = int(task_indices[i])
thr = thresholds[task_idx] threshold = thresholds[task_idx]
indicators[i] = 1 if float(advantages[i]) >= thr else 0 indicators[i] = 1 if float(advantages[i]) >= threshold else 0
if force_intervention_positive: if force_intervention_positive:
intervention_mask = interventions.astype(np.float32) > 0.5 intervention_mask = interventions.astype(np.float32) > 0.5

查看文件

@ -52,7 +52,9 @@ def update_policy(
if grad_clip_norm > 0: if grad_clip_norm > 0:
grad_norm = accelerator.clip_grad_norm_(policy.parameters(), grad_clip_norm) grad_norm = accelerator.clip_grad_norm_(policy.parameters(), grad_clip_norm)
else: else:
grad_norm = torch.nn.utils.clip_grad_norm_(policy.parameters(), float("inf"), error_if_nonfinite=False) grad_norm = torch.nn.utils.clip_grad_norm_(
policy.parameters(), float("inf"), error_if_nonfinite=False
)
with lock if lock is not None else nullcontext(): with lock if lock is not None else nullcontext():
optimizer.step() optimizer.step()

查看文件

@ -14,7 +14,6 @@
"""Human-in-loop recording helpers used by `lerobot_record.py`.""" """Human-in-loop recording helpers used by `lerobot_record.py`."""
import logging
from collections import deque from collections import deque
from concurrent.futures import ThreadPoolExecutor from concurrent.futures import ThreadPoolExecutor
from copy import deepcopy from copy import deepcopy
@ -53,7 +52,9 @@ def _get_torch_rng_state(device: torch.device) -> tuple[torch.Tensor, torch.Tens
return cpu_state, cuda_state return cpu_state, cuda_state
def _set_torch_rng_state(device: torch.device, cpu_state: torch.Tensor, cuda_state: torch.Tensor | None) -> None: def _set_torch_rng_state(
device: torch.device, cpu_state: torch.Tensor, cuda_state: torch.Tensor | None
) -> None:
torch.set_rng_state(cpu_state) torch.set_rng_state(cpu_state)
if device.type == "cuda" and cuda_state is not None: if device.type == "cuda" and cuda_state is not None:
torch.cuda.set_rng_state(cuda_state, device) torch.cuda.set_rng_state(cuda_state, device)

查看文件

@ -35,10 +35,10 @@ from lerobot.processor import (
) )
from lerobot.robots import Robot from lerobot.robots import Robot
from lerobot.scripts.recording_hil import ( from lerobot.scripts.recording_hil import (
ACPInferenceConfig,
INTERVENTION_STATE_ACTIVE, INTERVENTION_STATE_ACTIVE,
INTERVENTION_STATE_POLICY, INTERVENTION_STATE_POLICY,
INTERVENTION_STATE_RELEASE, INTERVENTION_STATE_RELEASE,
ACPInferenceConfig,
PolicySyncDualArmExecutor, PolicySyncDualArmExecutor,
_capture_policy_runtime_state, _capture_policy_runtime_state,
_predict_policy_action_with_acp_inference, _predict_policy_action_with_acp_inference,
@ -155,8 +155,8 @@ def record_loop(
action_feature_names = list(robot.action_features.keys()) action_feature_names = list(robot.action_features.keys())
else: else:
action_feature_names = list(robot.action_features) action_feature_names = list(robot.action_features)
zero_policy_action = {name: 0.0 for name in action_feature_names} zero_policy_action = dict.fromkeys(action_feature_names, 0.0)
has_teleop = isinstance(teleop, Teleoperator) or isinstance(teleop, list) has_teleop = isinstance(teleop, (Teleoperator, list))
intervention_enabled = intervention_state_machine_enabled and policy is not None and has_teleop intervention_enabled = intervention_state_machine_enabled and policy is not None and has_teleop
intervention_state = INTERVENTION_STATE_POLICY intervention_state = INTERVENTION_STATE_POLICY
last_teleop_action: RobotAction | None = None last_teleop_action: RobotAction | None = None
@ -282,22 +282,26 @@ def record_loop(
# Get action from policy and/or teleop # Get action from policy and/or teleop
act_processed_policy: RobotAction | None = None act_processed_policy: RobotAction | None = None
act_processed_teleop: RobotAction | None = None act_processed_teleop: RobotAction | None = None
if policy is not None and preprocessor is not None and postprocessor is not None: if (
if not (intervention_enabled and intervention_state == INTERVENTION_STATE_ACTIVE): policy is not None
policy_action = _predict_policy_action_with_acp_inference( and preprocessor is not None
observation_frame=observation_frame, and postprocessor is not None
policy=policy, and not (intervention_enabled and intervention_state == INTERVENTION_STATE_ACTIVE)
device=get_safe_torch_device(policy.config.device), ):
preprocessor=preprocessor, policy_action = _predict_policy_action_with_acp_inference(
postprocessor=postprocessor, observation_frame=observation_frame,
use_amp=policy.config.use_amp, policy=policy,
task=single_task, device=get_safe_torch_device(policy.config.device),
robot_type=robot.robot_type, preprocessor=preprocessor,
acp_inference=acp_inference, postprocessor=postprocessor,
cond_runtime_state=cond_policy_runtime_state, use_amp=policy.config.use_amp,
uncond_runtime_state=uncond_policy_runtime_state, task=single_task,
) robot_type=robot.robot_type,
act_processed_policy = make_robot_action(policy_action, dataset.features) acp_inference=acp_inference,
cond_runtime_state=cond_policy_runtime_state,
uncond_runtime_state=uncond_policy_runtime_state,
)
act_processed_policy = make_robot_action(policy_action, dataset.features)
if isinstance(teleop, Teleoperator): if isinstance(teleop, Teleoperator):
act = run_with_connection_retry("teleop.get_action", teleop.get_action) act = run_with_connection_retry("teleop.get_action", teleop.get_action)
@ -325,7 +329,9 @@ def record_loop(
last_teleop_action = act_processed_teleop last_teleop_action = act_processed_teleop
teleop_fallback_warned = False teleop_fallback_warned = False
policy_action_for_storage = act_processed_policy if act_processed_policy is not None else zero_policy_action policy_action_for_storage = (
act_processed_policy if act_processed_policy is not None else zero_policy_action
)
is_intervention = 0.0 is_intervention = 0.0
if intervention_enabled and intervention_state == INTERVENTION_STATE_ACTIVE: if intervention_enabled and intervention_state == INTERVENTION_STATE_ACTIVE:
@ -367,12 +373,14 @@ def record_loop(
if policy_sync_executor is not None and selected_from_policy: if policy_sync_executor is not None and selected_from_policy:
_sent_action = run_with_connection_retry( _sent_action = run_with_connection_retry(
"policy_sync_executor.send_action", "policy_sync_executor.send_action",
lambda: policy_sync_executor.send_action(robot_action_to_send), lambda robot_action_to_send=robot_action_to_send: policy_sync_executor.send_action(
robot_action_to_send
),
) )
else: else:
_sent_action = run_with_connection_retry( _sent_action = run_with_connection_retry(
"robot.send_action", "robot.send_action",
lambda: robot.send_action(robot_action_to_send), lambda robot_action_to_send=robot_action_to_send: robot.send_action(robot_action_to_send),
) )
# Write to dataset # Write to dataset

查看文件

@ -116,15 +116,11 @@ class BiSOLeader(Teleoperator):
def send_feedback(self, feedback: dict[str, float]) -> None: def send_feedback(self, feedback: dict[str, float]) -> None:
# Remove "left_" prefix # Remove "left_" prefix
left_feedback = { left_feedback = {
key.removeprefix("left_"): value key.removeprefix("left_"): value for key, value in feedback.items() if key.startswith("left_")
for key, value in feedback.items()
if key.startswith("left_")
} }
# Remove "right_" prefix # Remove "right_" prefix
right_feedback = { right_feedback = {
key.removeprefix("right_"): value key.removeprefix("right_"): value for key, value in feedback.items() if key.startswith("right_")
for key, value in feedback.items()
if key.startswith("right_")
} }
self.left_arm.send_feedback(left_feedback) self.left_arm.send_feedback(left_feedback)

查看文件

@ -165,9 +165,7 @@ class SOLeader(Teleoperator):
# For phase-A dual-arm execution, reuse leader as a commanded arm by # For phase-A dual-arm execution, reuse leader as a commanded arm by
# mapping `{joint}.pos` to motor goal positions. # mapping `{joint}.pos` to motor goal positions.
self.set_manual_control(False) self.set_manual_control(False)
goal_pos = { goal_pos = {key.removesuffix(".pos"): val for key, val in feedback.items() if key.endswith(".pos")}
key.removesuffix(".pos"): val for key, val in feedback.items() if key.endswith(".pos")
}
if not goal_pos: if not goal_pos:
return return
self.bus.sync_write("Goal_Position", goal_pos) self.bus.sync_write("Goal_Position", goal_pos)

查看文件

@ -103,4 +103,3 @@ def resolve_collector_policy_id(
if intervention_enabled: if intervention_enabled:
return human_id if is_intervention else policy_id return human_id if is_intervention else policy_id
return policy_id if selected_from_policy else human_id return policy_id if selected_from_policy else human_id

查看文件

@ -1 +1 @@
#!/usr/bin/env python #!/usr/bin/env python

查看文件

@ -12,7 +12,7 @@ from typing import TYPE_CHECKING, Any
import numpy as np import numpy as np
import torch import torch
import torch.nn.functional as F import torch.nn.functional as functional
from huggingface_hub import hf_hub_download from huggingface_hub import hf_hub_download
from huggingface_hub.constants import SAFETENSORS_SINGLE_FILE from huggingface_hub.constants import SAFETENSORS_SINGLE_FILE
from huggingface_hub.errors import HfHubHTTPError from huggingface_hub.errors import HfHubHTTPError
@ -78,7 +78,7 @@ def project_values_to_bins(values: torch.Tensor, bin_centers: torch.Tensor) -> t
def expected_value_from_logits(logits: torch.Tensor, bin_centers: torch.Tensor) -> torch.Tensor: def expected_value_from_logits(logits: torch.Tensor, bin_centers: torch.Tensor) -> torch.Tensor:
probs = F.softmax(logits, dim=-1) probs = functional.softmax(logits, dim=-1)
return (probs * bin_centers).sum(dim=-1) return (probs * bin_centers).sum(dim=-1)
@ -198,13 +198,13 @@ def _resolve_image_size(image_processor: Any) -> tuple[int, int]:
return 384, 384 return 384, 384
def _resolve_norm_stats(image_processor: Any) -> tuple[tuple[float, float, float], tuple[float, float, float]]: def _resolve_norm_stats(
image_processor: Any,
) -> tuple[tuple[float, float, float], tuple[float, float, float]]:
mean_raw = getattr(image_processor, "image_mean", [0.5, 0.5, 0.5]) mean_raw = getattr(image_processor, "image_mean", [0.5, 0.5, 0.5])
std_raw = getattr(image_processor, "image_std", [0.5, 0.5, 0.5]) std_raw = getattr(image_processor, "image_std", [0.5, 0.5, 0.5])
if len(mean_raw) != 3 or len(std_raw) != 3: if len(mean_raw) != 3 or len(std_raw) != 3:
raise ValueError( raise ValueError(f"Expected RGB normalization stats of len=3, got mean={mean_raw} std={std_raw}.")
f"Expected RGB normalization stats of len=3, got mean={mean_raw} std={std_raw}."
)
mean = (float(mean_raw[0]), float(mean_raw[1]), float(mean_raw[2])) mean = (float(mean_raw[0]), float(mean_raw[1]), float(mean_raw[2]))
std = (float(std_raw[0]), float(std_raw[1]), float(std_raw[2])) std = (float(std_raw[0]), float(std_raw[1]), float(std_raw[2]))
if any(v <= 0 for v in std): if any(v <= 0 for v in std):
@ -367,7 +367,7 @@ class Pistar06Model(nn.Module):
flat_images = images.view(bsize * num_cameras, *images.shape[2:]) flat_images = images.view(bsize * num_cameras, *images.shape[2:])
if flat_images.shape[-2:] != self.image_resolution: if flat_images.shape[-2:] != self.image_resolution:
flat_images = F.interpolate( flat_images = functional.interpolate(
flat_images, flat_images,
size=self.image_resolution, size=self.image_resolution,
mode="bilinear", mode="bilinear",
@ -435,7 +435,9 @@ class Pistar06Model(nn.Module):
language_context = torch.no_grad() if self.cfg.freeze_language_model else nullcontext() language_context = torch.no_grad() if self.cfg.freeze_language_model else nullcontext()
with language_context: with language_context:
language_features = self._encode_language(input_ids=input_ids, attention_mask=language_mask.long()) language_features = self._encode_language(
input_ids=input_ids, attention_mask=language_mask.long()
)
feature_dtype = torch.float32 feature_dtype = torch.float32
image_features = image_features.to(dtype=feature_dtype) image_features = image_features.to(dtype=feature_dtype)
@ -445,7 +447,9 @@ class Pistar06Model(nn.Module):
camera_token_mask = image_attention_mask.unsqueeze(-1).to(dtype=image_tokens.dtype) camera_token_mask = image_attention_mask.unsqueeze(-1).to(dtype=image_tokens.dtype)
image_tokens = image_tokens * camera_token_mask image_tokens = image_tokens * camera_token_mask
camera_denominator = image_attention_mask.sum(dim=1, keepdim=True).to(dtype=image_tokens.dtype).clamp_min(1.0) camera_denominator = (
image_attention_mask.sum(dim=1, keepdim=True).to(dtype=image_tokens.dtype).clamp_min(1.0)
)
image_pooled = image_tokens.sum(dim=1) / camera_denominator image_pooled = image_tokens.sum(dim=1) / camera_denominator
language_token = self.language_projector(language_features) language_token = self.language_projector(language_features)
@ -743,7 +747,7 @@ class Pistar06Policy(PreTrainedPolicy):
bin_centers = self.bin_centers.to(device=device) bin_centers = self.bin_centers.to(device=device)
soft_target = project_values_to_bins(value_target, bin_centers) soft_target = project_values_to_bins(value_target, bin_centers)
log_probs = F.log_softmax(logits, dim=-1) log_probs = functional.log_softmax(logits, dim=-1)
per_sample_loss = -(soft_target * log_probs).sum(dim=-1) per_sample_loss = -(soft_target * log_probs).sum(dim=-1)
sample_weight = None sample_weight = None
@ -769,13 +773,12 @@ class Pistar06Policy(PreTrainedPolicy):
pred_value = expected_value_from_logits(logits, bin_centers) pred_value = expected_value_from_logits(logits, bin_centers)
value_mae = (pred_value - value_target).abs().mean() value_mae = (pred_value - value_target).abs().mean()
if reduction == "none": loss = per_sample_loss if reduction == "none" else per_sample_loss.mean()
loss = per_sample_loss
else:
loss = per_sample_loss.mean()
loss_dict = { loss_dict = {
"loss": float(loss.mean().detach().item()) if reduction == "none" else float(loss.detach().item()), "loss": float(loss.mean().detach().item())
if reduction == "none"
else float(loss.detach().item()),
"value_mae": float(value_mae.detach().item()), "value_mae": float(value_mae.detach().item()),
} }
if sample_weight is not None: if sample_weight is not None:

查看文件

@ -8,7 +8,7 @@ from typing import Any
import numpy as np import numpy as np
import torch import torch
import torch.nn.functional as F import torch.nn.functional as functional
from torch import Tensor from torch import Tensor
from lerobot.configs.types import PipelineFeatureType, PolicyFeature from lerobot.configs.types import PipelineFeatureType, PolicyFeature
@ -39,7 +39,7 @@ PISTAR06_IMAGE_MASK_KEY = "observation.pistar06.image_attention_mask"
def _pad_last_dim(vector: Tensor, new_dim: int) -> Tensor: def _pad_last_dim(vector: Tensor, new_dim: int) -> Tensor:
if vector.shape[-1] >= new_dim: if vector.shape[-1] >= new_dim:
return vector return vector
return F.pad(vector, (0, new_dim - vector.shape[-1])) return functional.pad(vector, (0, new_dim - vector.shape[-1]))
@ProcessorStepRegistry.register(name="pistar06_prepare_task_prompt") @ProcessorStepRegistry.register(name="pistar06_prepare_task_prompt")
@ -216,7 +216,7 @@ def make_pistar06_pre_post_processors(
]: ]:
camera_features = list(config.camera_features) camera_features = list(config.camera_features)
if not camera_features: if not camera_features:
camera_features = [k for k in (config.input_features or {}).keys() if k.startswith(OBS_IMAGES)] camera_features = [k for k in (config.input_features or {}) if k.startswith(OBS_IMAGES)]
input_steps: list[ProcessorStep] = [ input_steps: list[ProcessorStep] = [
RenameObservationsProcessorStep(rename_map={}), RenameObservationsProcessorStep(rename_map={}),

查看文件

@ -729,7 +729,9 @@ def test_delete_episodes_preserves_extra_episode_metadata(tmp_path, empty_lerobo
) )
assert "episode_success" in new_dataset.meta.episodes.column_names assert "episode_success" in new_dataset.meta.episodes.column_names
assert [new_dataset.meta.episodes[i]["episode_success"] for i in range(new_dataset.meta.total_episodes)] == [ assert [
new_dataset.meta.episodes[i]["episode_success"] for i in range(new_dataset.meta.total_episodes)
] == [
"success", "success",
"success", "success",
] ]

查看文件

@ -15,9 +15,8 @@
# limitations under the License. # limitations under the License.
import json import json
from unittest.mock import MagicMock, patch
from collections import deque from collections import deque
from unittest.mock import MagicMock, patch
import numpy as np import numpy as np
import pytest import pytest
@ -25,12 +24,12 @@ import torch
from lerobot.datasets.lerobot_dataset import LeRobotDataset from lerobot.datasets.lerobot_dataset import LeRobotDataset
from lerobot.scripts.lerobot_calibrate import CalibrateConfig, calibrate from lerobot.scripts.lerobot_calibrate import CalibrateConfig, calibrate
from lerobot.scripts.lerobot_human_inloop_record import human_inloop_record
from lerobot.scripts.lerobot_human_inloop_record import ( from lerobot.scripts.lerobot_human_inloop_record import (
_HumanInloopFailureResetController, _HumanInloopFailureResetController,
_load_failure_reset_pose, _load_failure_reset_pose,
_save_failure_reset_pose, _save_failure_reset_pose,
_slow_reset_all_arms_to_pose, _slow_reset_all_arms_to_pose,
human_inloop_record,
) )
from lerobot.scripts.lerobot_record import ( from lerobot.scripts.lerobot_record import (
ACPInferenceConfig, ACPInferenceConfig,
@ -39,8 +38,8 @@ from lerobot.scripts.lerobot_record import (
RecordConfig, RecordConfig,
_capture_policy_runtime_state, _capture_policy_runtime_state,
_predict_policy_action_with_acp_inference, _predict_policy_action_with_acp_inference,
record_loop,
record, record,
record_loop,
) )
from lerobot.scripts.lerobot_replay import DatasetReplayConfig, ReplayConfig, replay from lerobot.scripts.lerobot_replay import DatasetReplayConfig, ReplayConfig, replay
from lerobot.scripts.lerobot_teleoperate import TeleoperateConfig, teleoperate from lerobot.scripts.lerobot_teleoperate import TeleoperateConfig, teleoperate
@ -211,9 +210,7 @@ def test_record_loop_sets_leader_manual_control_during_reset():
def test_save_and_load_failure_reset_pose(tmp_path): def test_save_and_load_failure_reset_pose(tmp_path):
robot = MockRobot( robot = MockRobot(MockRobotConfig(n_motors=2, random_values=False, static_values=[12.5, -3.0]))
MockRobotConfig(n_motors=2, random_values=False, static_values=[12.5, -3.0])
)
robot.connect() robot.connect()
pose_path = tmp_path / "failure_reset_pose.json" pose_path = tmp_path / "failure_reset_pose.json"

查看文件

@ -9,7 +9,9 @@ def test_contiguous_distributed_eval_sampler_has_contiguous_shards_with_tail_pad
shards: list[list[int]] = [] shards: list[list[int]] = []
for rank in range(world_size): for rank in range(world_size):
sampler = ContiguousDistributedEvalSampler(dataset_size=dataset_size, num_replicas=world_size, rank=rank) sampler = ContiguousDistributedEvalSampler(
dataset_size=dataset_size, num_replicas=world_size, rank=rank
)
shard = list(iter(sampler)) shard = list(iter(sampler))
shards.append(shard) shards.append(shard)
assert len(shard) == len(sampler) == 4 assert len(shard) == len(sampler) == 4
@ -30,7 +32,9 @@ def test_contiguous_distributed_eval_sampler_handles_more_ranks_than_samples():
shards: list[list[int]] = [] shards: list[list[int]] = []
for rank in range(world_size): for rank in range(world_size):
sampler = ContiguousDistributedEvalSampler(dataset_size=dataset_size, num_replicas=world_size, rank=rank) sampler = ContiguousDistributedEvalSampler(
dataset_size=dataset_size, num_replicas=world_size, rank=rank
)
shard = list(iter(sampler)) shard = list(iter(sampler))
shards.append(shard) shards.append(shard)
assert len(shard) == len(sampler) == 1 assert len(shard) == len(sampler) == 1

查看文件

@ -13,8 +13,7 @@ from lerobot.configs.value import (
ValueInferencePipelineConfig, ValueInferencePipelineConfig,
ValueInferenceVizConfig, ValueInferenceVizConfig,
) )
from lerobot.scripts import lerobot_value_infer as value_infer from lerobot.scripts import lerobot_value_infer as value_infer, value_infer_viz
from lerobot.scripts import value_infer_viz
class _FakeHFDataset: class _FakeHFDataset:

查看文件

@ -10,8 +10,6 @@ from torch import nn
import lerobot.processor.tokenizer_processor as tokenizer_processor import lerobot.processor.tokenizer_processor as tokenizer_processor
import lerobot.values.pistar06.modeling_pistar06 as pistar06_modeling import lerobot.values.pistar06.modeling_pistar06 as pistar06_modeling
import lerobot.values.pistar06.processor_pistar06 as pistar06_processor
from lerobot.configs.value_train import ValueTargetsConfig
from lerobot.utils.constants import OBS_LANGUAGE_ATTENTION_MASK, OBS_LANGUAGE_TOKENS, OBS_STATE from lerobot.utils.constants import OBS_LANGUAGE_ATTENTION_MASK, OBS_LANGUAGE_TOKENS, OBS_STATE
from lerobot.values.pistar06.configuration_pistar06 import Pistar06Config from lerobot.values.pistar06.configuration_pistar06 import Pistar06Config
from lerobot.values.pistar06.modeling_pistar06 import ( from lerobot.values.pistar06.modeling_pistar06 import (
@ -95,10 +93,7 @@ class _DummyAutoModel:
@classmethod @classmethod
def from_pretrained(cls, repo_id, **kwargs): def from_pretrained(cls, repo_id, **kwargs):
output_loading_info = bool(kwargs.pop("output_loading_info", False)) output_loading_info = bool(kwargs.pop("output_loading_info", False))
if "siglip" in repo_id.lower(): model = _DummyVisionModel() if "siglip" in repo_id.lower() else _DummyLanguageModel()
model = _DummyVisionModel()
else:
model = _DummyLanguageModel()
if output_loading_info: if output_loading_info:
return model, {"missing_keys": [], "unexpected_keys": [], "mismatched_keys": []} return model, {"missing_keys": [], "unexpected_keys": [], "mismatched_keys": []}
return model return model