Search public pages, research tools, and SAVRN solutions.
fastwam-droid-1p3b is an open-weight model for robotics from Angelzhang, released under Apache License 2.0. Its published files total 28.7 GB.
Action-conditioned robot world model (FastWAM MoT: Wan2.1-T2V-1.3B video branch + action DiT), trained on DreamZero-DROID. This is the PhyPO run droidphypo1p3b (runs/droidwan211p3b1e-4/2026-09-0801-06-31) at step 100,000.
By Angelzhang, published under apache-2.0, revision 8d80356a9c14.
Action-conditioned robot world model (FastWAM MoT: Wan2.1-T2V-1.3B video branch + action DiT), trained on DreamZero-DROID. This is the PhyPO run droidphypo1p3b (runs/droidwan211p3b1e-4/2026-09-0801-06-31) at step 100,000. The video DiT was initialized from the PhyPO/GRPO-finetuned Wan2.1-T2V-1.3B (Wan-AI-PhyPO-1.3B), and the action expert from ActionDiTphypoWan211p3Balphascale1024hdim.pt. This is not the stock Wan2.1 baseline (droid1p3bscratch). - step100000.pt — model weights (3.9 GB). Dict with keys mot, proprioencoder, step, torchdtype. - datasetstats.json — action/proprio normalization stats; required for inference. - The same checkpoint is also stored at…
Action-conditioned robot world model (FastWAM MoT: Wan2.1-T2V-1.3B video branch + action DiT),
trained on DreamZero-DROID. This is the PhyPO run droid_phypo_1p3b
(runs/droid_wan21_1p3b_1e-4/2026-09-08_01-06-31) at step 100,000.
The video DiT was initialized from the PhyPO/GRPO-finetuned Wan2.1-T2V-1.3B
(Wan-AI-PhyPO-1.3B), and the action expert from
ActionDiT_phypo_Wan21_1p3B_alphascale_1024hdim.pt.
This is not the stock Wan2.1 baseline (droid_1p3b_scratch).
| init | PhyPO Wan2.1-T2V-1.3B (30 layers, hidden 1536, 12 heads) + PhyPO ActionDiT |
| step | 100,000 (run max_steps was 200,000) |
| params | video expert 1.42B + action expert 0.64B, bfloat16 |
| data | DreamZero-DROID, 14,748,517 frames, 3 cameras (exterior_image_1_left, exterior_image_2_left, wrist_image_left) |
| proprio / action dim | 7 (cartesian + gripper) |
| train | 4× GPU, DeepSpeed ZeRO-1, bf16, grad clip 1.0, lr 1e-4 cosine |
| text conditioning | cached T5 embeds (wan21t2v13b), context_len 128 |
step_100000.pt — model weights (3.9 GB). Dict with keys mot, proprio_encoder, step, torch_dtype.dataset_stats.json — action/proprio normalization stats; required for inference.weights/2026-09-08_01-06-31/step_100000.pt.import torch
ck = torch.load("step_100000.pt", map_location="cpu", weights_only=True)
model.mot.load_state_dict(ck["mot"])
model.proprio_encoder.load_state_dict(ck["proprio_encoder"])
11 files, 28.7 GB in total. The weights are 7 files totalling 28.7 GB in pt, safetensors.
| File | Type | Size | SHA-256 |
|---|---|---|---|
| ActionDiT_linear_interp_Wan21_14B_alphascale_1024hdim.pt | Weights | 4.1 GB | a3d33b465e2a |
| ActionDiT_linear_interp_Wan21_14B_noalpha_1024hdim.pt | Weights | 8.1 GB | c4d3fdb55c21 |
| ActionDiT_linear_interp_Wan21_1p3B_alphascale_1024hdim.pt | Weights | 1.3 GB | 05b609e9f429 |
| ActionDiT_phypo_Wan21_1p3B_alphascale_1024hdim.pt | Weights | 1.3 GB | 4d89bbf73c43 |
| Wan-AI-PhyPO-1.3B/diffusion_pytorch_model.safetensors | Weights | 5.7 GB | 824dfbbef8bc |
| step_100000.pt | Weights | 4.1 GB | 72b5c30454d8 |
| weights/2026-09-08_01-06-31/step_100000.pt | Weights | 4.1 GB | 72b5c30454d8 |
| dataset_stats.json | Configuration | 47.7 KB | — |
| weights/2026-09-08_01-06-31/dataset_stats.json | Configuration | 47.7 KB | — |
| README.md | Documentation | 1.8 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
Released by Angelzhang through its official repository on Hugging Face. Read the license.
| Precision | Weights in memory |
|---|---|
| As published | 28.7 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Yes. fastwam-droid-1p3b is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
LIBERO 4in1（liberospatial / liberoobject / liberogoal / libero10）共 53.19 GB 的 Wan2.2-VAE 编码 latent 缓存： 训练 LIBERO policy（action head / VLA）时直接读取 latent 缓存，避免重复 VAE 编码。 - 窗口模式: windowed（--windowed） - 输出:.pt 文件，每 episode 一个 - dataset（推荐）: https://huggingface.co/datasets/MangoGoes/libero4in1wan2.2vaelatentdataset - model（本仓库）: https://huggingface.co/MangoGoes/libero4in1wan2.2vaelatentcosmosstyle
Wan2.2-TI2V-5B video DiT + 48-joint action head, trainingmode=joint. The base is suhyeok's finalized B-method recipe: a teacher-forced (sigma=0.25) self-EMA teacher plus an iBOT prototype loss at L18 L18, gamma=0.01, two-view. On top of it the 3 PAST cond latent frames are pooled into one motion frame before a chosen block. These are NOT the surrogate ctxpool runs. The surrogate line (older base, pd8 x GA1) lives in hmkang/wamctxpoolxattn and hmkang/wamctxpoolavg. Do not compare across the two sets. Geometry: 4-latin (numframesin=25, numframesout=41, fdf 2) = 4 cond + 2 future latent slots, 96 tokens per latent frame, 576 tokens per row. Effective batch 16 clips x GA 2 x 2 views = 64 rows…
weighted/imatrix quants of https://huggingface.co/DeepCybo/PhysBrain1.5-8B For a convenient overview and download list, visit our model page for this model. static quants are available at https://huggingface.co/mradermacher/PhysBrain1.5-8B-GGUF This is a vision model - mmproj files (if any) will be in the static repository. If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files. (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) Here is a handy graph by ikawrakow comparing some lower-quality quant And here are Artefact2's thoughts on the matter…
static quants of https://huggingface.co/MirilAI/Miril-DroneVLM-2B-2 For a convenient overview and download list, visit our model page for this model. weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion. If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files. (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) Here is a handy graph by ikawrakow comparing some lower-quality quant…
Own lineage: PPO from scratch (nb1, commit 46c48b1) continued as nb2 and nb4 (commit fac88cc), checkpoint 850 (1.328 s); evolution strategies on the weights against the Arena sprint scorer (0.952 s); then the network frozen and wrapped in a step counter carried in the headroll action that adds joint-target and forward-command offsets from about 0.5 m to the line, found by CMA-ES on the event's official seeds: 0.940 s. No other entry's weights used. Simulation only. A perpetual policy for the microduck (61-D observation, 14 actions, 50 Hz). Runs until told otherwise — a gait, loaded into a policy slot. The observation normalizer is baked into policy.onnx; feed raw observations. manifest.json…
EgoSteer-3B-RealMan is the generalist policy from EgoSteer: A Full-Stack System Towards Steerable Dexterous Manipulation from Egocentric Videos. It is a world-model-enhanced Vision-Language-Action (VLA) policy built on a Qwen3-VL backbone with a flow-matching action expert, post-trained from EgoSteer-3B-Base on real-world demonstrations collected on the RealMan robot. This is a robot-ready dual-camera (head + chest) checkpoint, corresponding to EgoSteer-9.6K-DG in the paper. For the base model to fine-tune on your own data, see EgoSteer-3B-Base. Our full-stack system integrates EgoSmith (data pipeline), Robot Stack (deployment), and EgoSteer (policy) to learn from 9.6k hours of large-scale…