Instructions to use rsj2003/StreamingWAM-DiT4DiT-LIBERO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use rsj2003/StreamingWAM-DiT4DiT-LIBERO with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 2,556 Bytes
81802d6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 | {
"date": "2026-11-12",
"scope": "Release smoke tests; the full 10,030-episode results are archived from the original runs, not rerun for packaging.",
"environment": {
"python": "3.10",
"torch": "2.7.1+cu128",
"precision": "bfloat16"
},
"cpu_contract_tests": {
"passed": 25
},
"existing_fastwam_unit_tests": {
"passed": 21
},
"strict_loading_and_inference": {
"joint": {
"method": "joint",
"strict_tensors": 1739,
"replans": 2,
"commands": 8,
"all_finite": true,
"peak_cuda_gib": 21.302031993865967,
"scope": "synthetic observation smoke test, not a benchmark evaluation"
},
"streaming": {
"method": "streaming",
"strict_tensors": 1739,
"replans": 2,
"commands": 8,
"all_finite": true,
"peak_cuda_gib": 21.45729684829712,
"scope": "synthetic observation smoke test, not a benchmark evaluation"
}
},
"training": {
"joint": {
"optimizer_updates": 1,
"world_size": 4,
"global_batch_size": 64,
"cosine_schedule_steps": 20000,
"completed_and_saved": true,
"action_loss": 0.002836579689756036,
"video_loss": 0.04728880152106285,
"total_loss": 0.05012538284063339
},
"streaming": {
"optimizer_updates": 1,
"world_size": 4,
"global_batch_size": 64,
"cosine_schedule_steps": 20000,
"completed_and_saved": true,
"action_loss": 0.0999070480465889,
"video_loss": 0.4279455542564392,
"total_loss": 0.5278525948524475
}
},
"simulator": {
"episodes_per_method": 1,
"episode_id": "libero_10_00000",
"joint": {
"steps": 700,
"success": false
},
"streaming": {
"steps": 326,
"success": true
},
"note": "Both public entrypoints completed without runtime errors. Outcomes match the corresponding archived case; action trajectories are not bitwise identical. This is not a new estimate of benchmark success."
},
"data_adapter": {
"fixture": "IPEC spatial episode 0",
"tested_frame_indices": [
0,
20,
60,
109
],
"comparison": "Public v2.1 importer and preparation vs. original v3 training input",
"checks": {
"joint": {
"samples": 4,
"actions_states_masks_identical": true,
"max_rgb_abs_difference": 0.0
},
"streaming": {
"samples": 4,
"actions_states_masks_identical": true,
"max_rgb_abs_difference": 0.0
}
}
},
"code_commit": "bb4c2cf"
}
|