Cheat sheet

All of the LeRobot commands in one place. If you forgot how to use a specific command or want to learn about a new one you can do it here.

For all of the commands listed below remember to change the ports/names/ids to your own values!

Another great way to look at all the commands and get them configured for your specific setup is to use this Jupyter Notebook.

Setup and installation

For installation please look at LeRobot Installation.

Useful tools

Find port

Use this to identify which serial ports your robots are connected to. Follow the instructions in your terminal: you will be asked to unplug the USB cable and press Enter. The script will then detect and print the correct serial port for that robot.

lerobot-find-port

Find cameras

Quickly find camera indices and verify their output. This command prints camera information to the terminal and saves test frames from each detected camera to outputs/captured_images

lerobot-find-cameras

Calibration

In most cases you will need to perform calibration just once for each robot and teleoperation device. Before performing the calibration make sure that all the joints are roughly in the middle position.

lerobot-calibrate \
    --robot.type=so101_follower \
    --robot.port=/dev/ttyACM0 \
    --robot.id=my_follower_arm

Make sure that you use the same IDs used during calibration later for the other scripts. That’s how LeRobot finds the calibration files.

Teleoperation

Teleoperating with two cameras and displaying the data with Rerun.

lerobot-teleoperate \
    --robot.type=so101_follower \
    --robot.port=/dev/ttyACM0 \
    --robot.id=my_follower_arm \
    --robot.cameras="{ top: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 30}, wrist: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30} }" \
    --teleop.type=so101_leader \
    --teleop.port=/dev/ttyACM1 \
    --teleop.id=my_leader_arm \
    --display_data=true

Recording a dataset

The dataset is automatically uploaded to the Hugging Face Hub under repo_id. Make sure you are logged in to your HF account with the CLI (you can get a token from https://huggingface.co/settings/tokens), and store your username in HF_USER, which the commands below use:

lerobot-record \
    --robot.type=so101_follower \
    --robot.port=/dev/ttyACM0 \
    --robot.id=my_follower_arm \
    --robot.cameras="{ top: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 30}, wrist: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30} }" \
    --teleop.type=so101_leader \
    --teleop.port=/dev/ttyACM1 \
    --teleop.id=my_leader_arm \
    --dataset.repo_id=${HF_USER}/so101_dataset_test \
    --dataset.num_episodes=30 \
    --dataset.single_task="put the red brick in a bowl" \
    --dataset.streaming_encoding=true \
    --display_data=true

While collecting the dataset you can control the process with your keyboard:

• Press Right Arrow (→) or n: Save episode and move to the next.
• Press Left Arrow (←) or r: Delete current episode and retry.
• Press Escape (ESC) or q: Stop, encode videos, and upload.

lerobot-record adds the date and time of the session to the end of --dataset.repo_id, so ${HF_USER}/so101_dataset_test is saved as ${HF_USER}/so101_dataset_test_YYYYMMDD_HHMMSS. Find the full name on your Hugging Face profile or with ls ~/.cache/huggingface/lerobot/${HF_USER}, and use it in the training commands below. Add --dataset.no_stamp=true to keep the exact name.

Recording depth

Intel RealSense cameras (type: intelrealsense) record a depth stream when you set use_depth: true. Depth is quantized to 12-bit codes and stored as its own video.

lerobot-record \
    ... \
    --robot.cameras="{ head: {type: intelrealsense, serial_number_or_name: \"0123456789\", width: 640, height: 480, fps: 30, use_depth: true} }" \
    --dataset.repo_id=${HF_USER}/so101_depth_test \
    --dataset.single_task="put the red brick in a bowl" \
    --dataset.depth_encoder.depth_min=0.01 \
    --dataset.depth_encoder.depth_max=10.0 \
    --dataset.depth_encoder.shift=0.0 \
    --dataset.depth_encoder.use_log=true

Video encoding parameters

RGB and depth streams are encoded independently via the --dataset.rgb_encoder.* and --dataset.depth_encoder.* keys.

lerobot-record \
    ... \
    --dataset.rgb_encoder.vcodec=h264 \
    --dataset.rgb_encoder.pix_fmt=yuv420p \
    --dataset.rgb_encoder.crf=23 \
    --dataset.depth_encoder.vcodec=hevc \
    --dataset.depth_encoder.extra_options='{"x265-params": "lossless=1"}'

Training

Depending on your hardware training the policy might take a few hours. That’s how you train simple ACT policy:

lerobot-train \
    --dataset.repo_id=${HF_USER}/so101_dataset_test_YYYYMMDD_HHMMSS \
    --policy.type=act \
    --output_dir=outputs/train/act_so101_test \
    --job_name=act_so101_test \
    --policy.device=cuda \
    --trackio.enable=true \
    --policy.repo_id=${HF_USER}/policy_test \
    --steps=20000

• Common policy types: act, diffusion, multi_task_dit, smolvla, pi0, pi05, pi0_fast, groot, xvla. Each policy has its own page under Policies in the sidebar, and the Compute & Hardware Guide shows which ones fit your GPU.
• Devices: cuda (NVIDIA), mps (Apple Silicon), cpu

--policy.type creates a new policy of that type and trains it from scratch on your dataset (some policies still start from pretrained vision or language backbones, as set in their config). To fine-tune an existing model instead, pass --policy.path with a local checkpoint or a Hub repo id: the policy type and config are read from it, so --policy.type is not needed.

lerobot-train \
    --dataset.repo_id=${HF_USER}/so101_dataset_test_YYYYMMDD_HHMMSS \
    --policy.path=username/the_policy_to_finetune \
    --policy.device=cuda \
    --policy.repo_id=${HF_USER}/policy_finetuned_test \
    --output_dir=outputs/train/finetuned_so101_test \
    --steps=20000

No local GPU? Add --job.target=<flavor> (e.g. a10g-small) to either command and lerobot-train runs it on Hugging Face Jobs instead — it uploads a local-only dataset for you and pushes the trained model. List flavors with hf jobs hardware. To track the job’s metrics, add --trackio.enable=true --trackio.space_id=<user>/<space>: the pod’s local Trackio database does not survive the job.

To resume, point --config_path at a checkpoint and add --resume=true. It accepts a local path or a Hub repo id (the latest checkpoint is fetched), and works locally or on a job by adding --job.target=<flavor>:

lerobot-train --config_path=${HF_USER}/policy_test --resume=true --job.target=a10g-small

Inference

Inference means running the trained policy/model on a robot. For that we use lerobot-rollout. You will need to provide a path to your policy. It can be a local path or a Hugging Face repo id, for example ${HF_USER}/policy_test from the training command above. Use the same port, --robot.id, camera configuration and task description as when collecting the dataset. --duration is in seconds; if unspecified, the policy runs until you stop it.

On v0.5.x and earlier, lerobot-rollout does not exist and you need to use lerobot-record instead. More information here.

lerobot-rollout \
    --strategy.type=base \
    --policy.path=${HF_USER}/policy_test \
    --robot.type=so101_follower \
    --robot.port=/dev/ttyACM0 \
    --robot.id=my_follower_arm \
    --robot.cameras="{ top: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 30}, wrist: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30} }" \
    --task="put the red brick in a bowl" \
    --duration=60