phospho-app/ACT_BBOX-ball20251109b-2jmikjwxbu

act model - 🧪 phosphobot training pipeline

• Dataset: phospho-app/ball20251109b_bboxes
• Wandb run id: None

Error Traceback

We faced an issue while training your model.

Training process failed with exit code 2:
[--scheduler.num_cycles str]
[--scheduler.num_warmup_steps str]
[--scheduler.num_decay_steps str] [--scheduler.peak_lr str]
[--scheduler.decay_lr str] [--eval str]
[--eval.n_episodes str] [--eval.batch_size str]
[--eval.use_async_envs str] [--wandb str] [--wandb.enable str]
[--wandb.disable_artifact str] [--wandb.project str]
[--wandb.entity str] [--wandb.notes str] [--wandb.run_id str]
[--wandb.mode str]
train.py: error: unrecognized arguments: --learning_rate=0.0001 --weight_decay=1e-06 --grad_clip_norm=1 --chunk_length=10 --action_horizon=10 --augmentations={'color_jitter': 0.1, 'gaussian_noise': 0.01, 'random_resized_crop': 0.1, 'horizontal_flip': False} --early_stopping={'enabled': True, 'patience': 10} --log_interval=100 --action_smoothing=True --multi_camera=True --mixed_precision=True

Training parameters

{
  "batch_size": 32,
  "steps": 30000,
  "save_freq": 5000,
  "target_detection_instruction": "red ball, green cup",
  "image_key": "observation.images.main",
  "image_keys_to_keep": [
    "observation.images.main",
    "observation.images.secondary_0"
  ],
  "learning_rate": 0.0001,
  "weight_decay": 1e-6,
  "grad_clip_norm": 1,
  "chunk_length": 10,
  "action_horizon": 10,
  "augmentations": {
    "color_jitter": 0.1,
    "gaussian_noise": 0.01,
    "random_resized_crop": 0.1,
    "horizontal_flip": false
  },
  "early_stopping": {
    "enabled": true,
    "patience": 10
  },
  "log_interval": 100,
  "action_smoothing": true,
  "multi_camera": true,
  "mixed_precision": true
}

📖 Get Started: docs.phospho.ai

🤖 Get your robot: robots.phospho.ai