THETA Bench
Collection
THETA Bench simulation policies (3,003 segments), selected real G1 adaptations (91 demos), and training data. • 15 items • Updated
THETA Bench FastWAM: Simulation training, 3,003 segments
The validated final policy checkpoint is available in this repository.
This is a THETA training-result repository. It does not substitute an upstream pretrained policy for a THETA-trained checkpoint.
| Setting | Value |
|---|---|
| Training stage | Simulation training, 3,003 segments |
| Target optimizer updates | 40000 |
| Per-GPU batch / GPUs / global batch | 16 / 8 / 128 |
| Gradient accumulation | 1 |
| Conditions per global batch | 18 |
| Dataset revision | 8b2cd31e107b64cb13f812ea217a63a20845c78a |
The simulation pool contains 1,200 successful L1/L2 demonstrations and 1,803 extracted L0 prefixes, spanning 18 conditions. The 3,003 segments are not 3,003 independent demonstrations.
Use the model's native THETA adapter and model-specific dependencies. This repository does not claim compatibility with arbitrary Transformers or simulation loaders. No evaluation score is claimed by checkpoint publication.
Native FastWAM weights contain mot and proprio_encoder. Use THETA _fastwam_server with --config-dir configs --task theta_g1 --dataset-stats run/dataset_stats.json and the final weight. Set THETA_FASTWAM_ACTION_DIT to the external initialization file matching inference_dependencies.json. Pre-cache its declared Wan2.2, T5/VAE and tokenizer dependencies. The native constructor loads those assets before trained weights, and native mot loading uses strict=False. This release does not establish standalone/offline loading or a new native inference qualification.
Training uses independent model optimizers and shared GPU execution through MPS. Publication is performed by a CPU uploader after final checkpoint validation.