Assessment

RoArm-M3-AI-Kit UART

{{{name2}}}

{{{name3}}}

{{{name4}}}

{{{name5}}}

{{{name6}}}

Primary Attribute Category: {{{userDefinedInfo}}}: {{{userdefinedvalue}}} Brand: Waveshare

Website International: {{{website_en}}} Chinese: {{{website_cn}}}

Onboard Interfaces UART

Related Products

LeRobot Tutorial Catalog

• 1. Install Environment
• 2. Configure Parameters
• 3. Remote Control Operation
• 4. Record Dataset
• 5. Visualize Dataset
• 6. Replay Dataset
• 7. Training
• 8. Assessment

8. Assessment

• Note: The following operations are performed inside the container
• Note: You need to navigate to the lerobot directory to execute the following code

8.1 Inference Model Preparation

8.1.1 Compression

After training, the location where the inference model is generated, take "/home/ws/lerobot/outputs/train/act_roarm_m3_test" as an example:

tar czvf act_roarm_m3_test.tar.gz outputs/train/act_roarm_m3_test/checkpoints/100000

Next, select the file and download it to your local device.

8.1.2 Decompression

Select the file and upload it to the "/home/ws/lerobot" directory of the inference environment.

Place the packaged dataset in the inference container, go to the root directory, and decompress it to "/home/ws/lerobot".

cd /home/ws/lerobot && tar xzvf act_roarm_m3_test.tar.gz

Enter the /home/ws/lerobot directory and create a link for the model:

ln -sfn 100000 outputs/train/act_roarm_m3_test/checkpoints/last

8.2 Record Assessment Dataset

Use the function in lerobot/scripts/control_robot.py but with strategy checkpoints as input. For example, run the following command to record 10 assessment sets:

python lerobot/scripts/control_robot.py \
  --robot.type=roarm_m3 \
  --control.type=record \
  --control.fps=30 \
  --control.single_task="Grasp a block and put it in the bin." \
  --control.repo_id=${HF_USER}/eval_act_roarm_m3_test \
  --control.tags='["tutorial"]' \
  --control.warmup_time_s=5 \
  --control.episode_time_s=30 \
  --control.reset_time_s=30 \
  --control.num_episodes=10 \
  --control.push_to_hub=true \
  --control.policy.path=outputs/train/act_roarm_m3_test/checkpoints/last/pretrained_model

Follow the official tutorial to learn about a more in-depth tutorial on controlling a real robot with LeRobot.