dddddsds31341/Long-WAM-Wash-Cup-Epoch-1

Long-WAM Wash Cup — Epoch 1

Evaluation weights from epoch 1 of 10, optimizer step 1453, of the independent wash-cup fine-tune. The run starts from Efficient-Large-Model/Long-WAM-YAM; it does not continue from another task's fine-tune.

Task instruction

Hold and lift the cup with one gripper while the other cleans its interior with the bottle brush, return the brush, then place the cup upside down on the drying rack.

Training setup

• Dataset: yuyingyinya/wash_cup, revision dcf90aa38d98e9b6c9f71afe0cfeeb3f50ed74cb.
• 100 demonstrations, 69,728 frames, 30 Hz; all demonstrations are used for training.
• Four GPUs, batch size 12 per GPU, gradient accumulation 1, global batch 48.
• Constant learning rate 1e-5, AdamW, weight decay 0.01; BF16 and DeepSpeed ZeRO-1.
• 1,453 optimizer updates per epoch; the full run targets 14,530 updates.
• 14-dimensional absolute joint targets with original gripper units, ordered left 6 joints + gripper, right 6 joints + gripper.
• Task-specific z-score statistics and the official [-5, 5] normalized-value clipping.
• Three RGB cameras: top_head, hand_left, hand_right; processed video layout 384×320, P48 history, 32-step action chunk.

Included files

• model.pt: joint video/action model and proprio encoder, stored step and dtype. No optimizer or scheduler state.
• config.yaml: the resolved configuration captured by the training run.
• dataset_stats.json: this task's normalization statistics; use these for preprocessing and action denormalization.
• code.zip: code-only SIP release v29 used for the training run, based on NVlabs/LongLive@cfa4e91eec59429d583326500bf5535c9ce6eec7 with the SIP changes included.
• requirements-sip.txt: recorded package versions for the SIP environment.

Loading and evaluation

Extract code.zip and install the included Long-WAM checkout. Use its longwam.inference.load_run_config and load_model with the explicit model.pt path, expected_step=1453 and strict=True. Model configuration and normalization statistics must come from this same repository.

The original training configuration retains SIP filesystem paths. For a different machine, set the model-assets root via LONGWAM_MODEL_ROOT and update data/statistics/cache paths for the intended evaluation entry point. Wan2.2 VAE, UMT5 text encoder and tokenizer assets are required separately. This repository is not a standalone Transformers from_pretrained model.

For deployment, use a matching robot adapter and the original 14D action/gripper conventions. This checkpoint has passed CPU file-loading checks; no closed-loop robot success rate or held-out validation result is reported. Training loss is not a robot-task success metric. The eval-only file cannot restore optimizer state for exact training continuation.

Online training record: https://forge.coreweave.com/wandb/bofang/long-wam/runs/3n9pgoq2

Base model and its published license metadata: https://huggingface.co/Efficient-Large-Model/Long-WAM-YAM