chore(dependencies): update uv.lock

Co-authored-by: Steven Palma <imstevenpmwork@ieee.org>

docs/source/_toctree.yaml.old

@@ -1,172 +0,0 @@
-- sections:
-  - local: index
-    title: LeRobot
-  - local: installation
-    title: Installation
-  - local: cheat-sheet
-    title: Cheat sheet
-  title: Get started
-- sections:
-  - local: il_robots
-    title: Imitation Learning for Robots
-  - local: bring_your_own_policies
-    title: Adding a Policy
-  - local: integrate_hardware
-    title: Bring Your Own Hardware
-  - local: hilserl
-    title: Train a Robot with RL
-  - local: hilserl_sim
-    title: Train RL in Simulation
-  - local: multi_gpu_training
-    title: Multi GPU training
-  - local: hil_data_collection
-    title: Human In the Loop Data Collection
-  - local: peft_training
-    title: Training with PEFT (e.g., LoRA)
-  - local: rename_map
-    title: Using Rename Map and Empty Cameras
-  title: "Tutorials"
-- sections:
-  - local: hardware_guide
-    title: Compute Hardware Guide
-  - local: torch_accelerators
-    title: PyTorch accelerators
-  title: "Compute & Hardware"
-- sections:
-  - local: lerobot-dataset-v3
-    title: Using LeRobotDataset
-  - local: porting_datasets_v3
-    title: Porting Large Datasets
-  - local: using_dataset_tools
-    title: Using the Dataset Tools
-  - local: language_and_recipes
-    title: Language Columns and Recipes
-  - local: tools
-    title: Tools
-  - local: video_encoding_parameters
-    title: Video encoding parameters
-  - local: streaming_video_encoding
-    title: Streaming Video Encoding
-  title: "Datasets"
-- sections:
-  - local: act
-    title: ACT
-  - local: smolvla
-    title: SmolVLA
-  - local: pi0
-    title: π₀ (Pi0)
-  - local: pi0fast
-    title: π₀-FAST (Pi0Fast)
-  - local: pi05
-    title: π₀.₅ (Pi05)
-  - local: eo1
-    title: EO-1
-  - local: groot
-    title: NVIDIA GR00T N1.5
-  - local: xvla
-    title: X-VLA
-  - local: multi_task_dit
-    title: Multitask DiT Policy
-  - local: walloss
-    title: WALL-OSS
-  title: "Policies"
-- sections:
-  - local: sarm
-    title: SARM
-  title: "Reward Models"
-- sections:
-  - local: inference
-    title: Policy Deployment (lerobot-rollout)
-  - local: async
-    title: Use Async Inference
-  - local: rtc
-    title: Real-Time Chunking (RTC)
-  title: "Inference"
-- sections:
-  - local: envhub
-    title: Environments from the Hub
-  - local: envhub_leisaac
-    title: Control & Train Robots in Sim (LeIsaac)
-  title: "Simulation"
-- sections:
-  - local: adding_benchmarks
-    title: Adding a New Benchmark
-  - local: libero
-    title: LIBERO
-  - local: libero_plus
-    title: LIBERO-plus
-  - local: metaworld
-    title: Meta-World
-  - local: robotwin
-    title: RoboTwin 2.0
-  - local: robocasa
-    title: RoboCasa365
-  - local: robocerebra
-    title: RoboCerebra
-  - local: robomme
-    title: RoboMME
-  - local: envhub_isaaclab_arena
-    title: NVIDIA IsaacLab Arena Environments
-  - local: vlabench
-    title: VLABench
-  title: "Benchmarks"
-- sections:
-  - local: introduction_processors
-    title: Introduction to Robot Processors
-  - local: debug_processor_pipeline
-    title: Debug your processor pipeline
-  - local: implement_your_own_processor
-    title: Implement your own processor
-  - local: processors_robots_teleop
-    title: Processors for Robots and Teleoperators
-  - local: env_processor
-    title: Environment Processors
-  - local: action_representations
-    title: Action Representations
-  title: "Robot Processors"
-- sections:
-  - local: so101
-    title: SO-101
-  - local: so100
-    title: SO-100
-  - local: koch
-    title: Koch v1.1
-  - local: lekiwi
-    title: LeKiwi
-  - local: hope_jr
-    title: Hope Jr
-  - local: reachy2
-    title: Reachy 2
-  - local: unitree_g1
-    title: Unitree G1
-  - local: earthrover_mini_plus
-    title: Earth Rover Mini
-  - local: omx
-    title: OMX
-  - local: openarm
-    title: OpenArm
-  - local: rebot_b601
-    title: reBot B601-DM
-  title: "Robots"
-- sections:
-  - local: phone_teleop
-    title: Phone
-  title: "Teleoperators"
-- sections:
-  - local: cameras
-    title: Cameras
-  title: "Sensors"
-- sections:
-  - local: notebooks
-    title: Notebooks
-  - local: feetech
-    title: Updating Feetech Firmware
-  - local: damiao
-    title: Damiao Motors and CAN Bus
-  title: "Resources"
-- sections:
-  - local: contributing
-    title: Contribute to LeRobot
-  - local: backwardcomp
-    title: Backward compatibility
-  title: "About"

docs/source/_toctree.yml

@@ -1,214 +1,172 @@
-# LeRobot documentation table of contents
-#
-# Ordering principle: gentle onboarding first, advanced/custom work last.
-# Within each top-level section the same rule applies — concept/overview pages
-# before reference/per-item pages.
-#
-# Pages marked "NEW (to create)" do not yet exist as .mdx files; they are
-# placeholders for the redesign and must be authored before the docs build.
-
 - sections:
   - local: index
-    title: 🤗 LeRobot
+    title: LeRobot
-  - local: quickstart           # NEW (to create) — 15-min zero-to-trained-ACT path
-    title: Quickstart
   - local: installation
     title: Installation
-  - local: core_concepts        # NEW (to create) — datasets, policies, processors, robots, envs in one mental model
-    title: Core concepts
   - local: cheat-sheet
-    title: Command cheat sheet
+    title: Cheat sheet
   title: Get started
-
 - sections:
   - local: il_robots
-    title: Imitation learning end-to-end
+    title: Imitation Learning for Robots
+  - local: bring_your_own_policies
+    title: Adding a Policy
+  - local: integrate_hardware
+    title: Bring Your Own Hardware
+  - local: hilserl
+    title: Train a Robot with RL
+  - local: hilserl_sim
+    title: Train RL in Simulation
+  - local: multi_gpu_training
+    title: Multi GPU training
   - local: hil_data_collection
-    title: Human-in-the-loop data collection
+    title: Human In the Loop Data Collection
-  - local: inference
+  - local: peft_training
-    title: Deploying a trained policy
+    title: Training with PEFT (e.g., LoRA)
   - local: rename_map
-    title: Matching dataset keys to a policy (rename map)
+    title: Using Rename Map and Empty Cameras
-  title: Your first project
+  title: "Tutorials"
-
 - sections:
   - local: hardware_guide
-    title: Compute hardware guide
+    title: Compute Hardware Guide
   - local: torch_accelerators
     title: PyTorch accelerators
-  - local: multi_gpu_training
+  title: "Compute & Hardware"
-    title: Multi-GPU training
-  - local: peft_training
-    title: Parameter-efficient fine-tuning (LoRA)
-  title: Training
-
 - sections:
   - local: lerobot-dataset-v3
     title: Using LeRobotDataset
+  - local: porting_datasets_v3
+    title: Porting Large Datasets
   - local: using_dataset_tools
-    title: Dataset tools
+    title: Using the Dataset Tools
   - local: language_and_recipes
-    title: Language columns & recipes
+    title: Language Columns and Recipes
   - local: tools
-    title: Tool calls in datasets
+    title: Tools
   - local: video_encoding_parameters
     title: Video encoding parameters
   - local: streaming_video_encoding
-    title: Streaming video encoding
+    title: Streaming Video Encoding
-  - local: porting_datasets_v3
+  title: "Datasets"
-    title: Porting datasets to v3
-  title: Datasets
-
 - sections:
-  - local: policies_overview    # NEW (to create) — concept hub + "choose a policy" decision guide
+  - local: act
-    title: Choosing a policy
+    title: ACT
-  - sections:
+  - local: smolvla
-    - local: act
+    title: SmolVLA
-      title: ACT
+  - local: pi0
-    - local: smolvla
+    title: π₀ (Pi0)
-      title: SmolVLA
+  - local: pi0fast
-    - local: pi0
+    title: π₀-FAST (Pi0Fast)
-      title: π₀ (Pi0)
+  - local: pi05
-    - local: pi0fast
+    title: π₀.₅ (Pi05)
-      title: π₀-FAST
+  - local: eo1
-    - local: pi05
+    title: EO-1
-      title: π₀.₅ (Pi05)
+  - local: groot
-    - local: eo1
+    title: NVIDIA GR00T N1.5
-      title: EO-1
+  - local: xvla
-    - local: groot
+    title: X-VLA
-      title: NVIDIA GR00T N1.5
+  - local: multi_task_dit
-    - local: xvla
+    title: Multitask DiT Policy
-      title: X-VLA
+  - local: walloss
-    - local: walloss
+    title: WALL-OSS
-      title: WALL-OSS
+  title: "Policies"
-    - local: multi_task_dit
-      title: Multitask DiT
-    title: Policy reference
-  title: Policies
-
 - sections:
-  - local: async
-    title: Async inference
-  - local: rtc
-    title: Real-time chunking (RTC)
-  title: Real-time deployment
-
-- sections:
-  - local: hilserl
-    title: Train a robot with RL (HIL-SERL)
-  - local: hilserl_sim
-    title: Train RL in simulation
   - local: sarm
-    title: SARM reward model
+    title: SARM
-  title: Reinforcement learning
+  title: "Reward Models"
-
+- sections:
+  - local: inference
+    title: Policy Deployment (lerobot-rollout)
+  - local: async
+    title: Use Async Inference
+  - local: rtc
+    title: Real-Time Chunking (RTC)
+  title: "Inference"
 - sections:
   - local: envhub
     title: Environments from the Hub
   - local: envhub_leisaac
-    title: LeIsaac — control & train in sim
+    title: Control & Train Robots in Sim (LeIsaac)
+  title: "Simulation"
+- sections:
+  - local: adding_benchmarks
+    title: Adding a New Benchmark
+  - local: libero
+    title: LIBERO
+  - local: libero_plus
+    title: LIBERO-plus
+  - local: metaworld
+    title: Meta-World
+  - local: robotwin
+    title: RoboTwin 2.0
+  - local: robocasa
+    title: RoboCasa365
+  - local: robocerebra
+    title: RoboCerebra
+  - local: robomme
+    title: RoboMME
   - local: envhub_isaaclab_arena
-    title: NVIDIA IsaacLab Arena environments
+    title: NVIDIA IsaacLab Arena Environments
-  - sections:
+  - local: vlabench
-    - local: libero
+    title: VLABench
-      title: LIBERO
+  title: "Benchmarks"
-    - local: libero_plus
-      title: LIBERO-plus
-    - local: metaworld
-      title: Meta-World
-    - local: robotwin
-      title: RoboTwin 2.0
-    - local: robocasa
-      title: RoboCasa365
-    - local: robocerebra
-      title: RoboCerebra
-    - local: robomme
-      title: RoboMME
-    - local: vlabench
-      title: VLABench
-    title: Benchmark suites
-  title: Simulation & benchmarks
-
 - sections:
   - local: introduction_processors
-    title: Introduction to processors
+    title: Introduction to Robot Processors
-  - local: processors_robots_teleop
-    title: Processors for robots & teleoperators
-  - local: env_processor
-    title: Environment processors
-  - local: action_representations
-    title: Action representations
   - local: debug_processor_pipeline
-    title: Debugging a pipeline
+    title: Debug your processor pipeline
   - local: implement_your_own_processor
-    title: Implementing your own processor
+    title: Implement your own processor
-  title: Processors
+  - local: processors_robots_teleop
-
+    title: Processors for Robots and Teleoperators
+  - local: env_processor
+    title: Environment Processors
+  - local: action_representations
+    title: Action Representations
+  title: "Robot Processors"
 - sections:
-  - sections:
+  - local: so101
-    - local: so101
+    title: SO-101
-      title: SO-101
+  - local: so100
-    - local: so100
+    title: SO-100
-      title: SO-100
+  - local: koch
-    - local: koch
+    title: Koch v1.1
-      title: Koch v1.1
+  - local: lekiwi
-    - local: omx
+    title: LeKiwi
-      title: OMX
+  - local: hope_jr
-    - local: openarm
+    title: Hope Jr
-      title: OpenArm
+  - local: reachy2
-    title: Low-cost arms
+    title: Reachy 2
-  - sections:
+  - local: unitree_g1
-    - local: lekiwi
+    title: Unitree G1
-      title: LeKiwi
+  - local: earthrover_mini_plus
-    - local: earthrover_mini_plus
+    title: Earth Rover Mini
-      title: Earth Rover Mini
+  - local: omx
-    title: Mobile platforms
+    title: OMX
-  - sections:
+  - local: openarm
-    - local: hope_jr
+    title: OpenArm
-      title: Hope Jr
+  - local: rebot_b601
-    - local: reachy2
+    title: reBot B601-DM
-      title: Reachy 2
+  title: "Robots"
-    - local: unitree_g1
+- sections:
-      title: Unitree G1
+  - local: phone_teleop
-    title: Bimanual & humanoid
+    title: Phone
-  - sections:
+  title: "Teleoperators"
-    - local: rebot_b601
-      title: reBot B601-DM
-    title: Research & industrial
-  title: Supported robots
-
 - sections:
   - local: cameras
     title: Cameras
-  - local: phone_teleop
+  title: "Sensors"
-    title: Phone teleoperation
-  - local: feetech
-    title: Feetech firmware update
-  - local: damiao
-    title: Damiao motors & CAN bus
-  title: Sensors, teleop & motors
-
 - sections:
-  - local: integrate_hardware
-    title: Bring your own hardware
-  - local: bring_your_own_policies
-    title: Add a new policy
-  - local: adding_benchmarks
-    title: Add a new benchmark
-  title: Extend LeRobot
-
-- sections:
-  - local: troubleshooting       # NEW (to create) — common errors: USB, calibration drift, CUDA OOM, video decoding…
-    title: Troubleshooting & FAQ
-  - local: glossary              # NEW (to create) — episode, action chunk, leader/follower, teleop, processor…
-    title: Glossary
   - local: notebooks
-    title: Example notebooks
+    title: Notebooks
-  - local: backwardcomp
+  - local: feetech
-    title: Backward compatibility
+    title: Updating Feetech Firmware
-  title: Reference
+  - local: damiao
-
+    title: Damiao Motors and CAN Bus
+  title: "Resources"
 - sections:
   - local: contributing
-    title: Contributing to LeRobot
+    title: Contribute to LeRobot
-  title: About
+  - local: backwardcomp
+    title: Backward compatibility
+  title: "About"

docs/source/quickstart.mdx

@@ -1,219 +0,0 @@
-# Quickstart
-
-This is the **shortest path** from an unboxed SO-101 to a policy that drives your own robot. Every step is copy-paste; replace the **`<placeholders>`** with the values for your setup.
-
-By the end you will have:
-
-- A calibrated SO-101 leader + follower pair.
-- A dataset of 30 episodes pushed to the Hugging Face Hub.
-- A trained ACT policy (~20k steps) running on your robot via `lerobot-rollout`.
-
-> [!NOTE]
-> **How long will this take?**
-> Recording 30 episodes is roughly 30–60 minutes of teleoperation. Training ACT for 20k steps takes ~1.5h on an A100, a few hours on a laptop RTX 3060, longer on Apple Silicon (`mps`). The commands themselves are quick — most of the wall-clock is data collection and training.
-
-> [!TIP]
-> If you only want to **understand the codebase** or **train on an existing dataset without hardware**, this page isn't for you. Read [Core concepts](./core_concepts) first, then jump to [Imitation learning end-to-end](./il_robots).
-
----
-
-## Before you start
-
-You need:
-
-- An **assembled SO-101 leader + follower pair**. If your robot is not assembled yet, follow the [SO-101 assembly guide](./so101) and come back here.
-- **One or two cameras** (USB webcam works fine).
-- A **CUDA GPU with ≥ 6 GB VRAM** (ACT is light — a laptop RTX 3060 works). Apple Silicon (`mps`) and CPU are supported but slower. See the [compute hardware guide](./hardware_guide) for sizing.
-- A **Hugging Face account** — datasets and the trained policy will be pushed to your Hub.
-
-If any of the above is missing, fix it first; the rest of the page assumes it.
-
----
-
-## Step 1 — Install LeRobot
-
-Follow the full [Installation Guide](./installation) for environment setup, then add the SO-101 motor stack and log in to the Hub:
-
-```bash
-pip install 'lerobot[feetech]'
-git lfs install && git lfs pull
-hf auth login                 # paste a token from https://huggingface.co/settings/tokens
-```
-
-Sanity check — the CLI entry points should be available:
-
-```bash
-lerobot-find-port --help
-```
-
----
-
-## Step 2 — Identify USB ports and motor IDs
-
-Plug **only the follower arm** in (USB + power) and run:
-
-```bash
-lerobot-find-port
-```
-
-When prompted, unplug it and press Enter. Note the printed port — that's your `<FOLLOWER_PORT>`. Repeat with only the **leader arm** plugged in to get `<LEADER_PORT>`.
-
-> [!TIP]
-> On Linux, USB ports look like `/dev/ttyACM0`; on macOS like `/dev/tty.usbmodem...`. On Linux you may need `sudo chmod 666 /dev/ttyACM0` to grant access.
-
-If your motors are brand-new (or repurposed), set their IDs and baudrate **once per arm**:
-
-```bash
-lerobot-setup-motors --robot.type=so101_follower --robot.port=<FOLLOWER_PORT>
-lerobot-setup-motors --teleop.type=so101_leader  --teleop.port=<LEADER_PORT>
-```
-
-The script walks you through connecting motors one at a time. Full details: [SO-101 → Configure the motors](./so101#configure-the-motors).
-
----
-
-## Step 3 — Calibrate
-
-Center every joint roughly in the middle of its range, then run:
-
-```bash
-lerobot-calibrate \
-    --robot.type=so101_follower \
-    --robot.port=<FOLLOWER_PORT> \
-    --robot.id=my_follower
-
-lerobot-calibrate \
-    --teleop.type=so101_leader \
-    --teleop.port=<LEADER_PORT> \
-    --teleop.id=my_leader
-```
-
-After pressing Enter, sweep each joint through its full range of motion, then press Enter again to finish.
-
-> [!WARNING]
-> The `--robot.id` / `--teleop.id` values (`my_follower`, `my_leader`) become the **calibration keys**. Reuse the same IDs in every later command — that's how LeRobot finds the calibration on disk.
-
-Watch the [calibration video](./so101#calibrate) if anything is unclear.
-
----
-
-## Step 4 — Teleoperate (sanity check, no recording)
-
-Before recording anything, confirm the leader drives the follower correctly:
-
-```bash
-lerobot-teleoperate \
-    --robot.type=so101_follower \
-    --robot.port=<FOLLOWER_PORT> \
-    --robot.id=my_follower \
-    --robot.cameras="{ top: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30} }" \
-    --teleop.type=so101_leader \
-    --teleop.port=<LEADER_PORT> \
-    --teleop.id=my_leader \
-    --display_data=true
-```
-
-A Rerun window should open showing the camera feed and joint angles. Move the leader — the follower should mirror it in real time. If it doesn't, see [Troubleshooting & FAQ](./troubleshooting).
-
-Don't know which camera index is which? Run `lerobot-find-cameras` — it saves a frame from each detected camera so you can pick the right one.
-
----
-
-## Step 5 — Record a dataset (30 episodes)
-
-Now record demonstrations. Pick a short, repeatable task (e.g. *"put the red brick in the bowl"*). The dataset is pushed to the Hub under your username:
-
-```bash
-export HF_USER=<your-hf-username>
-
-lerobot-record \
-    --robot.type=so101_follower \
-    --robot.port=<FOLLOWER_PORT> \
-    --robot.id=my_follower \
-    --robot.cameras="{ top: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30}, wrist: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 30} }" \
-    --teleop.type=so101_leader \
-    --teleop.port=<LEADER_PORT> \
-    --teleop.id=my_leader \
-    --dataset.repo_id=${HF_USER}/so101_quickstart \
-    --dataset.num_episodes=30 \
-    --dataset.single_task="Put the red brick in the bowl" \
-    --dataset.streaming_encoding=true \
-    --display_data=true
-```
-
-**Keyboard controls during recording:**
-
-- **`→` (Right Arrow)** — save the current episode and move to the next.
-- **`←` (Left Arrow)** — discard the current episode and retry.
-- **`Esc`** — stop, encode videos, and upload to the Hub.
-
-> [!TIP]
-> **Quality beats quantity.** 30 clean, varied episodes (different brick positions, lighting, camera shake) train a much better policy than 100 identical ones. Move the object around. Vary your speed slightly.
-
-When you're done, your dataset lives at `https://huggingface.co/datasets/${HF_USER}/so101_quickstart`. You can preview it in the browser. For deeper recording options (resume, multiple tasks, custom processors), see [Imitation learning end-to-end → Record](./il_robots#record-a-dataset).
-
----
-
-## Step 6 — Train ACT
-
-ACT (Action Chunking Transformer) is the right default for a first run — small, fast, and works well on 30 episodes.
-
-```bash
-lerobot-train \
-    --dataset.repo_id=${HF_USER}/so101_quickstart \
-    --policy.type=act \
-    --output_dir=outputs/train/act_so101_quickstart \
-    --job_name=act_so101_quickstart \
-    --policy.device=cuda \
-    --policy.repo_id=${HF_USER}/act_so101_quickstart \
-    --steps=20000 \
-    --wandb.enable=true
-```
-
-A few notes:
-
-- Replace `--policy.device=cuda` with `mps` on Apple Silicon, or `cpu` if you have no GPU (very slow — not recommended for a real run).
-- `--wandb.enable=true` is optional. If you use it, run `wandb login` first. Otherwise drop the flag.
-- Checkpoints land in `outputs/train/act_so101_quickstart/checkpoints/`. The final model is also pushed to the Hub at the `--policy.repo_id` you specified.
-- To resume from an interruption: `lerobot-train --config_path=outputs/train/act_so101_quickstart/checkpoints/last/pretrained_model/train_config.json --resume=true`.
-
-> [!TIP]
-> **No GPU locally?** Train on Google Colab using the [ACT notebook](./notebooks#training-act), or rent a GPU via [Hugging Face Jobs](./il_robots#train-using-hugging-face-jobs) — pay-as-you-go, no setup.
-
-For why ACT is the default and when to switch to SmolVLA, Pi0, or another policy, see [Choosing a policy](./policies_overview).
-
----
-
-## Step 7 — Run your policy on the robot
-
-Deploy with `lerobot-rollout`. **Use the same camera layout you used while recording** — keys and resolutions must match.
-
-```bash
-lerobot-rollout \
-    --strategy.type=base \
-    --policy.path=${HF_USER}/act_so101_quickstart \
-    --robot.type=so101_follower \
-    --robot.port=<FOLLOWER_PORT> \
-    --robot.id=my_follower \
-    --robot.cameras="{ top: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30}, wrist: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 30} }" \
-    --task="Put the red brick in the bowl" \
-    --duration=60
-```
-
-`--duration` is in seconds — leave it off to run until you stop the script. You should see the follower arm move on its own, attempting the task.
-
-If observations from the robot use different keys than the policy expects, you'll need a [rename map](./rename_map). If latency matters, look at [async inference](./async) and [real-time chunking](./rtc).
-
----
-
-## You're done 🎉
-
-You now have a working IL pipeline end-to-end. From here, the natural next steps are:
-
-- **Improve the policy** — record more diverse episodes, train longer, or try a stronger model. See [Choosing a policy](./policies_overview).
-- **Go deeper on imitation learning** — [Imitation learning end-to-end](./il_robots) covers multi-camera setups, multi-task datasets, episode replay, evaluation, and Hugging Face Jobs.
-- **Try RL with a human in the loop** — [HIL-SERL](./hilserl) trains a policy that improves while you correct it.
-- **Use a different robot** — see [Supported robots](./so101) for low-cost arms, mobile platforms, bimanual, and humanoid.
-- **Build something new** — [Bring your own hardware](./integrate_hardware) and [Add a new policy](./bring_your_own_policies).
-
-Stuck on something? Check [Troubleshooting & FAQ](./troubleshooting), or ask on [Discord](https://discord.gg/s3KuuzsPFb).

src/lerobot/configs/parser.py

@@ -255,8 +255,7 @@ def extract_path_fields_from_config(config_path: str, path_fields: list[str]) ->
             remaining = config_data[field]
             if remaining:
                 _config_yaml_overrides[field] = _flatten_to_cli_args(remaining)
-            else:
+            del config_data[field]
-                del config_data[field]
             modified = True
 
     if not modified:
@@ -311,7 +310,13 @@ def wrap(config_path: Path | None = None) -> Callable[[F], F]:
                     cli_args = filter_arg("config_path", cli_args)
                     cfg = argtype.from_pretrained(config_path_cli, cli_args=cli_args)
                 else:
-                    cfg = draccus.parse(config_class=argtype, config_path=config_path, args=cli_args)
+                    if config_path_cli:
+                        cli_args = filter_arg("config_path", cli_args)
+                    cfg = draccus.parse(
+                        config_class=argtype,
+                        config_path=config_path_cli or config_path,
+                        args=cli_args,
+                    )
             response = fn(cfg, *args, **kwargs)
             return response
 

src/lerobot/policies/groot/eagle2_hg_model/modeling_eagle2_5_vl.py

@@ -60,6 +60,7 @@ class Eagle25VLPreTrainedModel(PreTrainedModel):
         "SiglipEncoderLayer",
     ]
     _skip_keys_device_placement = "past_key_values"
+    _supports_flash_attn = True
     _supports_flash_attn_2 = True
     _supports_cache_class = True
     _supports_static_cache = True

src/lerobot/policies/groot/eagle2_hg_model/processing_eagle2_5_vl.py

@@ -124,7 +124,6 @@ class Eagle25VLProcessor(ProcessorMixin):
         "videos_kwargs",
         "text_kwargs",
     ]
-    image_processor_class = "AutoImageProcessor"
     tokenizer_class = "AutoTokenizer"
 
     def __init__(

src/lerobot/policies/groot/processor_groot.py

@@ -206,7 +206,11 @@ def _build_eagle_processor(tokenizer_assets_repo: str = DEFAULT_TOKENIZER_ASSETS
             "Vendor files are copied during model creation. Create the policy/model first, "
             "or call ensure_eagle_cache_ready() before building processors."
         )
-    proc = AutoProcessor.from_pretrained(str(cache_dir), trust_remote_code=True, use_fast=True)
+    proc = AutoProcessor.from_pretrained(
+        str(cache_dir),
+        trust_remote_code=True,
+        fix_mistral_regex=False,
+    )
     proc.tokenizer.padding_side = "left"
     return proc
 

tests/test_yaml_policy_path.py

@@ -1,10 +1,14 @@
 """Tests for policy.path support in YAML config files (issue #2957)."""
 
 import json
+import sys
 import tempfile
+from dataclasses import dataclass, field
+from unittest.mock import patch
 
 import yaml
 
+from lerobot.configs import parser
 from lerobot.configs.parser import (
     _config_path_args,
     _config_yaml_overrides,
@@ -16,7 +20,8 @@ from lerobot.configs.parser import (
 
 
 def test_extract_path_fields_from_yaml():
-    """Test that policy.path is extracted from a YAML config and removed."""
+    """Test that policy.path is extracted from a YAML config and the policy block
+    is removed entirely (siblings are captured separately as cli_overrides)."""
     config = {
         "dataset": {"repo_id": "lerobot/pusht"},
         "policy": {"type": "smolvla", "path": "lerobot/smolvla_base", "push_to_hub": False},
@@ -26,26 +31,33 @@ def test_extract_path_fields_from_yaml():
         config_path = f.name
 
     _config_path_args.clear()
+    _config_yaml_overrides.clear()
     cleaned_path = extract_path_fields_from_config(config_path, ["policy"])
 
     # Path should be extracted and stored
     assert _config_path_args["policy"] == "lerobot/smolvla_base"
 
-    # Cleaned config should not have the path field
+    # Cleaned config should not have the policy block at all -- draccus must not
+    # try to decode it as PreTrainedConfig; the actual config comes from
+    # from_pretrained(path) with the captured overrides applied on top.
     with open(cleaned_path) as f:
         cleaned = yaml.safe_load(f)
-    assert "path" not in cleaned["policy"]
+    assert "policy" not in cleaned
-    assert cleaned["policy"]["type"] == "smolvla"
-    assert cleaned["policy"]["push_to_hub"] is False
 
     # Original dataset should be untouched
     assert cleaned["dataset"]["repo_id"] == "lerobot/pusht"
 
+    # Sibling overrides (excluding type/path) captured for from_pretrained.
+    overrides = get_yaml_overrides("policy")
+    assert any("push_to_hub=false" in o for o in overrides)
+
     _config_path_args.clear()
+    _config_yaml_overrides.clear()
 
 
 def test_extract_path_fields_from_json():
-    """Test that policy.path is extracted from a JSON config."""
+    """Test that policy.path is extracted from a JSON config and the policy
+    block is removed entirely."""
     config = {
         "policy": {"type": "act", "path": "some/local/path"},
     }
@@ -54,15 +66,17 @@ def test_extract_path_fields_from_json():
         config_path = f.name
 
     _config_path_args.clear()
+    _config_yaml_overrides.clear()
     cleaned_path = extract_path_fields_from_config(config_path, ["policy"])
 
     assert _config_path_args["policy"] == "some/local/path"
 
     with open(cleaned_path) as f:
         cleaned = json.load(f)
-    assert "path" not in cleaned["policy"]
+    assert "policy" not in cleaned
 
     _config_path_args.clear()
+    _config_yaml_overrides.clear()
 
 
 def test_extract_no_path_returns_original():
@@ -216,3 +230,91 @@ def test_flatten_nested_with_bools():
     args = _flatten_to_cli_args(d)
     assert "--optimizer.use_warmup=true" in args
     assert "--optimizer.lr=0.01" in args
+
+
+def test_extract_removes_field_with_siblings_and_no_type():
+    """Regression: when policy.path has siblings but no type:, the entire policy
+    block must still be removed from the cleaned config. Otherwise draccus tries
+    to decode the leftover dict as PreTrainedConfig and crashes on the missing
+    type discriminator.
+    """
+    config = {
+        "dataset": {"repo_id": "lerobot/pusht"},
+        "policy": {
+            "path": "lerobot/smolvla_base",
+            "n_action_steps": 10,
+            "dtype": "bfloat16",
+        },
+    }
+    with tempfile.NamedTemporaryFile(mode="w", suffix=".yaml", delete=False) as f:
+        yaml.dump(config, f)
+        config_path = f.name
+
+    _config_path_args.clear()
+    _config_yaml_overrides.clear()
+    cleaned_path = extract_path_fields_from_config(config_path, ["policy"])
+
+    with open(cleaned_path) as f:
+        cleaned = yaml.safe_load(f) or {}
+    assert "policy" not in cleaned, "policy block should be fully removed when path is present"
+    assert cleaned["dataset"]["repo_id"] == "lerobot/pusht"
+    assert _config_path_args["policy"] == "lerobot/smolvla_base"
+    overrides = get_yaml_overrides("policy")
+    assert any("n_action_steps=10" in o for o in overrides)
+    assert any("dtype=bfloat16" in o for o in overrides)
+
+    _config_path_args.clear()
+    _config_yaml_overrides.clear()
+
+
+@dataclass
+class _DummyNested:
+    foo: int = 0
+
+
+@dataclass
+class _DummyConfig:
+    nested: _DummyNested = field(default_factory=_DummyNested)
+    other: str = "default"
+
+    @classmethod
+    def __get_path_fields__(cls):
+        return ["nested"]
+
+
+def test_wrap_uses_cleaned_config_for_draccus_parse():
+    """Regression: wrap() updates config_path_cli to point at the cleaned temp
+    file but must propagate that to the draccus.parse fallback branch. Without
+    the fix, cli_args still contains --config_path=<original> and draccus reads
+    the original YAML with `path:` still in it, crashing on the unknown field.
+    """
+    config = {
+        "nested": {"path": "some/checkpoint", "foo": 42},
+        "other": "set-via-yaml",
+    }
+    with tempfile.NamedTemporaryFile(mode="w", suffix=".yaml", delete=False) as f:
+        yaml.dump(config, f)
+        config_path = f.name
+
+    _config_path_args.clear()
+    _config_yaml_overrides.clear()
+
+    captured: dict = {}
+
+    @parser.wrap()
+    def main(cfg: _DummyConfig) -> _DummyConfig:
+        captured["cfg"] = cfg
+        return cfg
+
+    with patch.object(sys, "argv", ["prog", f"--config_path={config_path}"]):
+        main()
+
+    assert captured["cfg"].other == "set-via-yaml"
+    assert _config_path_args["nested"] == "some/checkpoint"
+    # Cleaned config dropped `nested:` entirely; defaults stand for this wrapper
+    # class (a real PreTrainedConfig would now load the checkpoint and apply
+    # the captured yaml_overrides via from_pretrained()).
+    assert captured["cfg"].nested.foo == 0
+
+    _config_path_args.clear()
+    _config_yaml_overrides.clear()