=== Avatar Training (fire-and-forget) === Script: train_on_hf.py Started: 2026-03-12 17:49:21 UTC >>> [0/4] Raising file descriptor limit... ulimit -n = 1048576 >>> [1/4] Setting environment... HF_TOKEN set: 37 chars WANDB_API_KEY set: 0 chars >>> [2/4] Logging into HuggingFace... The token has not been saved to the git credentials helper. Pass `add_to_git_credential=True` in this function directly or `--add-to-git-credential` if using via `hf`CLI if you want to set the git credential as well. Token is valid (permission: write). The token `beta-pc` has been saved to /root/.cache/huggingface/stored_tokens Your token has been saved to /root/.cache/huggingface/token Login successful. Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. ⚠️ Warning: 'huggingface-cli login' is deprecated. Use 'hf auth login' instead. >>> [3/4] Pre-downloading pretrained model... No pretrained path to pre-download. GPUs: 1x NVIDIA GeForce RTX 4090 >>> [4/4] Launching accelerate with 1 GPUs... Script: train_on_hf.py Train args: --dataset alpha_bot_right_afternoon --steps 400000 --batch-size 12 --lr 1e-5 --preload --mixed-precision bf16 --compile --num-workers 8 --prefetch-factor 4 --save-freq 10000 --push-to-hub --wandb Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 [Avatar] Auto-tagged dataset with version: v3.0 [Avatar] Mixed precision: bf16 [Avatar] W&B key invalid or missing (0 chars, need 40+), skipping W&B ============================================================ Training args: mixed_precision=bf16 video_backend=pyav prefetch_factor=4 preload=True compile=True backbone=resnet18 lerobot_train \ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \ --dataset.video_backend=pyav \ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \ --policy.type=act \ --policy.chunk_size=100 \ --policy.n_action_steps=100 \ --policy.vision_backbone=resnet18 \ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \ --policy.dim_model=512 \ --policy.n_heads=8 \ --policy.dim_feedforward=3200 \ --policy.n_encoder_layers=4 \ --policy.n_decoder_layers=1 \ --policy.use_vae=true \ --policy.latent_dim=32 \ --policy.n_vae_encoder_layers=4 \ --policy.dropout=0.1 \ --policy.kl_weight=10.0 \ --policy.optimizer_lr=1e-05 \ --policy.optimizer_weight_decay=1e-4 \ --policy.optimizer_lr_backbone=1e-05 \ --batch_size=12 \ --steps=400000 \ --save_freq=10000 \ --log_freq=200 \ --save_checkpoint=true \ --seed=1000 \ --num_workers=8 \ --output_dir=outputs/train/act-vacuum-pick \ --policy.push_to_hub=true \ --policy.private=true ============================================================ [Avatar] cudnn.benchmark=True, tf32=True [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 (with per-step tags) [Avatar] torch.compile enabled (will compile on first forward pass) [Avatar] Overriding prefetch_factor=4 WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'. INFO 2026-03-12 17:49:32 ot_train.py:195 {'batch_size': 12, 'checkpoint_path': None, 'dataset': {'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'affine': {'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}, 'type': 'RandomAffine', 'weight': 1.0}, 'brightness': {'kwargs': {'brightness': [0.8, 1.2]}, 'type': 'ColorJitter', 'weight': 1.0}, 'contrast': {'kwargs': {'contrast': [0.8, 1.2]}, 'type': 'ColorJitter', 'weight': 1.0}, 'hue': {'kwargs': {'hue': [-0.05, 0.05]}, 'type': 'ColorJitter', 'weight': 1.0}, 'saturation': {'kwargs': {'saturation': [0.5, 1.5]}, 'type': 'ColorJitter', 'weight': 1.0}, 'sharpness': {'kwargs': {'sharpness': [0.5, 1.5]}, 'type': 'SharpnessJitter', 'weight': 1.0}}}, 'repo_id': 'avatarrobotics/alpha-bot-right-2026-03-11-afternoon', 'revision': None, 'root': None, 'streaming': False, 'use_imagenet_stats': True, 'video_backend': 'pyav'}, 'env': None, 'eval': {'batch_size': 50, 'n_episodes': 50, 'use_async_envs': False}, 'eval_freq': 20000, 'job_name': 'act', 'log_freq': 200, 'num_workers': 8, 'optimizer': {'betas': [0.9, 0.999], 'eps': 1e-08, 'grad_clip_norm': 10.0, 'lr': 1e-05, 'type': 'adamw', 'weight_decay': 0.0001}, 'output_dir': 'outputs/train/act-vacuum-pick', 'peft': None, 'policy': {'chunk_size': 100, 'device': 'cuda', 'dim_feedforward': 3200, 'dim_model': 512, 'dropout': 0.1, 'feedforward_activation': 'relu', 'input_features': {}, 'kl_weight': 10.0, 'latent_dim': 32, 'license': None, 'n_action_steps': 100, 'n_decoder_layers': 1, 'n_encoder_layers': 4, 'n_heads': 8, 'n_obs_steps': 1, 'n_vae_encoder_layers': 4, 'normalization_mapping': {'ACTION': , 'STATE': , 'VISUAL': }, 'optimizer_lr': 1e-05, 'optimizer_lr_backbone': 1e-05, 'optimizer_weight_decay': 0.0001, 'output_features': {}, 'pre_norm': False, 'pretrained_backbone_weights': 'ResNet18_Weights.IMAGENET1K_V1', 'pretrained_path': None, 'private': True, 'push_to_hub': True, 'replace_final_stride_with_dilation': False, 'repo_id': 'avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749', 'tags': None, 'temporal_ensemble_coeff': None, 'type': 'act', 'use_amp': False, 'use_peft': False, 'use_vae': True, 'vision_backbone': 'resnet18'}, 'rabc_epsilon': 1e-06, 'rabc_head_mode': 'sparse', 'rabc_kappa': 0.01, 'rabc_progress_path': None, 'rename_map': {}, 'resume': False, 'save_checkpoint': True, 'save_freq': 10000, 'scheduler': None, 'seed': 1000, 'steps': 400000, 'tolerance_s': 0.0001, 'use_policy_training_preset': True, 'use_rabc': False, 'wandb': {'disable_artifact': False, 'enable': False, 'entity': None, 'mode': None, 'notes': None, 'project': 'lerobot', 'run_id': None}} INFO 2026-03-12 17:49:32 ot_train.py:203 Logs will be saved locally. INFO 2026-03-12 17:49:32 ot_train.py:215 Creating dataset Fetching 7 files: 0%| | 0/7 [00:00>> Training finished with exit code 0 Ended: 2026-03-12 23:22:06 UTC >>> Saving training log to HuggingFace...