Chapter 15: ATEC2026 Lightweight Code
Quick Reference of Common LeHome Commands
Download Assets
bashhf download lehome/asset_challenge --repo-type dataset --local-dir Assets
Download and merge training set
bashhf download lehome/dataset_challenge_merged --repo-type dataset --local-dir Datasets/example
Train ACT
bashmkdir -p /root/gpufree-data/lehome-outputs/train/act_top_short
lerobot-train \
--config_path /root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/configs/train_act_every_embodied.yaml \
2>&1 | tee /root/gpufree-data/lehome-outputs/train/act_top_short/train.log
Train DP
bashmkdir -p /root/gpufree-data/lehome-outputs/train/dp_top_short
lerobot-train \
--config_path /root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/configs/train_dp_every_embodied.yaml \
2>&1 | tee /root/gpufree-data/lehome-outputs/train/dp_top_short/train.log
Evaluation of ACT
bashpython -m scripts.eval \
--policy_type lerobot \
--policy_path /root/gpufree-data/lehome-outputs/train/act_top_short/checkpoints/last/pretrained_model \
--dataset_root Datasets/example/top_short_merged \
--garment_type top_short \
--num_episodes 2 \
--enable_cameras \
--save_video \
--video_dir /root/gpufree-data/lehome-outputs/eval/act_top_short \
--device cpu
Evaluation of DP
bashpython -m scripts.eval \
--policy_type lerobot \
--policy_path /root/gpufree-data/lehome-outputs/train/dp_top_short/checkpoints/last/pretrained_model \
--dataset_root Datasets/example/top_short_merged \
--garment_type top_short \
--num_episodes 2 \
--enable_cameras \
--save_video \
--video_dir /root/gpufree-data/lehome-outputs/eval/dp_top_short \
--device cpu \
--policy_device cpu \
--policy_num_inference_steps 1
Parse training logs and create diagrams
bashpython /root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/scripts/plot_train_metrics.py \
--log_file /root/gpufree-data/lehome-outputs/train/act_top_short/train.log \
--out_dir /root/gpufree-data/lehome-outputs/plots/act_top_short \
--title "ACT Top-Short Training Metrics"
Competition Enhanced Version of ACT Long Training
bashcd /root/lehome-challenge
source .venv/bin/activate
export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
lerobot-train \
--config_path /root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/configs/train_act_competition_l40.yaml \
2>&1 | tee /root/gpufree-data/lehome-outputs/train/act_four_types_l40.log
Sampling GPU Metrics
bash/root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/scripts/sample_gpu_metrics.sh \
/root/gpufree-data/lehome-outputs/monitor/act_four_types_l40_gpu.csv \
10
Draw GPU curve
bashpython /root/gpufree-data/every-embodied/15-Challenge竞赛/LeHome/resources/scripts/plot_gpu_metrics.py \
--csv_file /root/gpufree-data/lehome-outputs/monitor/act_four_types_l40_gpu.csv \
--out_png /root/gpufree-data/lehome-outputs/plots/act_four_types_l40_live/gpu_metrics.png \
--title "L40 GPU Monitor During ACT Training"