lerobot_train_model_multi_gpu_wandb.sh
Humanoid/fmc3-robotics-main/projects/lerobot/demo_scripts/lerobot_train_model_multi_gpu_wandb.sh
export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
accelerate launch \
--mixed_precision=fp16 \
--multi_gpu \
--num_processes=2 \
$(which lerobot-train) \
--dataset.repo_id=${HF_USER}/pick_place_with_rgbd \
--policy.type=act \
--output_dir=outputs/train/act_multi_gpu/lerobot_model \
--job_name=pick_and_place_bottle_from_A_to_B \
--policy.device=cuda \
--policy.use_amp=true \
--policy.precision=fp16 \
--batch_size=16 \
--steps=50000 \
--optimizer.lr=2e-4 \
--wandb.enable=true \
--policy.repo_id=${HF_USER}/lerobot_pick_and_place_act_rgbd
lerobot_train_model_SmolVLAmulti_gpu_wandb.sh
Sau →lerobot_train_model_multi_gpu_wandb_diffusion.sh
Bài viết liên quan
run.py
run.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/deploy/run.py).
Đọc bài viết →utils.py
utils.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/deploy/utils.py).
Đọc bài viết →agent.py
agent.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/master/agents/agent.py).
Đọc bài viết →planner.py
planner.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/master/agents/planner.py).
Đọc bài viết →prompts.py
prompts.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/master/agents/prompts.py).
Đọc bài viết →run.py
run.py — python source code from the Humanoid learning materials (Humanoid/May_12/RoboOS/master/run.py).
Đọc bài viết →