# Open-HumanWAM: 50K Hours of Egocentric Video for Robot WAMs > Source: > Published: 2026-11-09 00:00:00+00:00 **Peking University's DAGroup published Open-HumanWAM**, an open recipe to pretrain **world–action models** on **fifty thousand hours** of egocentric human video and post-train them into robot policies that lead reported **RoboTwin 2.0** numbers among compared open checkpoints. The [arXiv paper](https://arxiv.org/abs/2611.06817?ref=explainx) ( **2611.06817** ), [project site](https://dagroup-pku.github.io/open-humanwam/?ref=explainx), [Hugging Face models](https://huggingface.co/collections/DAGroup-PKU/open-humanwam?ref=explainx), and [dataset card](https://huggingface.co/datasets/DAGroup-PKU/Open-HumanWAM-50K?ref=explainx) went live in early November 2026. If you follow embodied AI because [human egocentric video is the scalable manipulation prior](https://explainx.ai/blog/egocentric-data-collection-robots-explained-2026), this release is one of the first to bundle **corpus, checkpoints, training code, and benchmark scripts** under a single branding umbrella rather than a leaderboard screenshot alone. ## TL;DR | Question | Answer | |---|---| | Corpus | **Open-HumanWAM-50K** — curated egocentric video, hand motion + language | | Models | **HumanWAM-Vanilla** (human-only pretrain) and**HumanWAM-Goal** (human + robot, visual goals) | | Key sim result (post-trained) | Up to **~88.65%** average on**RoboTwin 2.0** Clean + Random per project tables | | Scaling claim | **Log-linear** link between pretraining hours and downstream success; diversity matters | | Open assets | [GitHub HumanNet branch](https://github.com/DAGroup-PKU/HumanNet/tree/open-humanscale?ref=explainx) , HF weights + dataset | | Builder angle | Reproduce scaling curves before betting your robot stack on proprietary human-video APIs | ## What the paper adds beyond another VLA headline The abstract on [Hugging Face Papers](https://huggingface.co/papers/2611.06817?ref=explainx) states the problem bluntly: raw human video scale does not automatically become generalizable robot policies without **curation** and a **training recipe**. Open-HumanWAM claims three contributions readers can verify if they download the assets: 1. **A 50k-hour corpus** with dense hand-motion and natural-language annotation—not a token YouTube scrape without labels. 2. **Systematic scaling experiments** in simulation that produce an empirical**log-linear law** between human-video hours and robot success after adaptation. 3. **Two model lines** —HumanWAM-Vanilla (human pretrain only) and HumanWAM-Goal (joint human–robot with visual goals)—with**open weights** and evaluation harnesses. HumanWAM-Vanilla, after robot post-training, is reported at **92.08%** success on RoboTwin 2.0 **Clean2Clean** and **68.88%** on **Clean2Random**, with additional **RoboDojo** metrics in the paper text (about **14.17%** mean success and **19.00** mean progress score in the HF summary). HumanWAM-Goal pushes the harder randomized split to **84.60%**, averaging **88.65%** across Clean and Random in the post-trained regime described on the project site. Those numbers compete with other 2026 world–action releases we have covered—[Verne V0.5](https://explainx.ai/blog/verne-robotics-v0-5-open-weight-world-action-model-8b-libero-99-3-robotwin-81-gated-license-claim-check-2026) reported **81.0% / 73.4%** clean/random on RoboTwin with an 8B checkpoint, and [SLA continual pretraining](https://explainx.ai/blog/sla-spatial-language-action-model-continual-pretraining-vla-robotics-arxiv-2611-08373-2026) reported **72.3%** Clean2Random. **Do not treat leaderboard ranks as eternal**—each paper uses slightly different post-training budgets—but Open-HumanWAM's selling point is that you can **download their checkpoints and rerun the same RoboTwin scripts**. ## HumanWAM-Vanilla: scaling human experience HumanWAM study protocol from egocentric pretrain through RoboTwin adaptation *Image: [PKU DAGroup](https://dagroup-pku.github.io/open-humanwam/?ref=explainx)* The [project microsite](https://dagroup-pku.github.io/open-humanwam/?ref=explainx) walks through a three-step protocol: 1. **Learn from people** — video, language, and hand motion in egocentric frames. 2. **Adapt to robots** — post-train on**50 RoboTwin tasks** with**50 clean demonstrations** each. 3. **Test generalization** — clean scenes, scene out-of-distribution, and randomized clutter. Scaling charts compare **100 → 10,000 → 50,000 hours** under both **isolated attention** and **inverse-dynamics attention** variants. The authors emphasize gains **outside training scenes**, not only memorizing tabletop layouts. Inverse-dynamics attention scaling from 100 to 50,000 hours of egocentric pretraining *Image: [PKU DAGroup](https://dagroup-pku.github.io/open-humanwam/?ref=explainx)* A second study splits **scaled-diversity** corpora (more tasks and scenes as hours grow) from **coverage-matched** corpora (more trajectories within fixed task/scene coverage). The site argues **diversity scaling** produces steeper transfer than repeating similar clips—a lesson for teams paying for [egocentric annotation APIs](https://explainx.ai/blog/perceptron-egocentric-1-0-api-9-dollars-video-hour-egohome-gpt-6-astra-one-eighth-cost-2026) without widening scene coverage. Scaled-diversity versus coverage-matched corpus scaling curves *Image: [PKU DAGroup](https://dagroup-pku.github.io/open-humanwam/?ref=explainx)* ## HumanWAM-Goal: visual goals and zero-shot behavior HumanWAM-Goal trains a **visual-goal-conditioned** world–action model on **hybrid human and robot data**. Before RoboTwin-specific post-training, the project lists **13.94%** average zero-shot success across Clean and Random, which they describe as leading among **successfully evaluated open-source zero-shot** checkpoints in their comparison table—not the same protocol as the post-trained **88.65%** figure. On real hardware, the site reports **20.0% → 30.5%** average success across twenty tasks when moving from **robot-only** pretraining to hybrids with **15.9k** and **45.9k hours** of human data, with robot experience fixed at **14,618 hours** in that ablation. Qualitative demos show **goal-conditioned in-context learning**: the policy keeps weights frozen while the visual target changes—useful for teams thinking about [Natura-1-style dexterous transfer](https://explainx.ai/blog/nature-will-natura-1-dexterous-hands-foundation-model-egocentric-human-video-2026) without collecting hundreds of new teleop demos per SKU. Long-horizon clips include **desk organization** (fourteen stages) and mobile tidying on **Astribot S1**, plus **cross-embodiment** tests where a **human-view goal image** guides a different robot embodiment after task post-training. Treat those videos as **authors' demonstrations** until independent labs reproduce them. ## How to try it this week **Read and cite** - Paper PDF: [arxiv.org/abs/2611.06817](https://arxiv.org/abs/2611.06817?ref=explainx) - Overview and films: [dagroup-pku.github.io/open-humanwam/](https://dagroup-pku.github.io/open-humanwam/?ref=explainx) **Download** - Models collection: `huggingface.co/collections/DAGroup-PKU/open-humanwam` - Dataset: `huggingface.co/datasets/DAGroup-PKU/Open-HumanWAM-50K` - Code: `github.com/DAGroup-PKU/HumanNet` branch`open-humanscale` **Evaluate** - Start in **RoboTwin 2.0** with their published post-training recipe before attempting real robots. - Log **Clean vs Random** separately; Random is where[Xiaomi U0-style synthesis](https://explainx.ai/blog/xiaomi-robotics-u0-world-foundation-model-embodied-synthesis-july-2026) and domain-randomization papers claim the real generalization win. **Compute expectations** Fifty thousand hours of video is not a laptop fine-tune. Plan cluster storage for the corpus subset you need, mirror HF artifacts regionally, and budget GPU time for world-model pretraining at the scales their curves imply—100-hour baselines are useful sanity checks, not the headline model. ## RoboTwin 2.0 and RoboDojo in plain language **RoboTwin 2.0** is a two-arm manipulation benchmark with **fifty tasks**, each trained from tens of clean demonstrations, then tested in **matched clean scenes** and **randomized scenes** where object poses, textures, and lighting shift. Papers report **Clean2Clean** and **Clean2Random** separately because randomized success collapses when policies memorize a single tabletop layout. Open-HumanWAM's **~80% vanilla / ~89% goal** averages are **post-trained** numbers: they assume you run the authors' adaptation stage on RoboTwin demonstrations after human-video pretraining. The **13.94% zero-shot** figure for HumanWAM-Goal is a different protocol—evaluating the pretrained checkpoint **before** task-specific post-training—so it should not be compared directly to the eighty-eight percent line in the same breath. **RoboDojo** metrics in the paper (mid-teens percent success in the HF abstract) act as a secondary stress test. Treat them as **additional evidence**, not the headline marketers will tweet. ## World–action models versus classic VLAs Classic **vision-language-action (VLA)** stacks map an image and instruction directly to joint targets or end-effector deltas. **World–action models (WAMs)** also model short-horizon **dynamics** or **video prediction** jointly with actions, which can help when contact-rich physics matter. HumanWAM's architecture diagrams on the project site emphasize **closed-loop simulation evaluation** during scaling studies—that is, the model is not judged only on next-token loss on human YouTube clips. For builders comparing stacks, our [LLM vs CLM vs decision models guide](https://explainx.ai/blog/llm-vs-clm-vs-decision-models-which-to-use-2026) is about language heads, not robot WAMs, but the same lesson applies: **match the evaluation harness to the deployment harness** before choosing a pretrained initialization. If you already run a **π₀-class diffusion policy** or a smaller **8B Verne checkpoint**, treat HumanWAM as a **pretraining initializer** experiment: swap the human-video encoder weights, keep your robot data collection pipeline fixed for a month, and measure **data efficiency** (demos needed to hit 70% on your task) rather than chasing leaderboard points on tasks you will never ship. ## What people are asking on Hugging Face The author comment on the [paper page](https://huggingface.co/papers/2611.06817?ref=explainx) summarizes HumanWAM-Vanilla at **92.08% / 68.88%** Clean/Random (**~80.48%** average) and HumanWAM-Goal at **92.70% / 84.60%** (**~88.65%** average) after post-training, plus **RoboTwin rank #1** language in their self-reported comparison. A community comment from **PolyHelperAI** notes the **log-linear** hours-to-success relationship and asks where the curve **flattens**—a fair open question the paper's appendices may address with larger held-out suites. Treat flattening points as **empirical**, not guaranteed log-linear forever. The automated **Librarian Bot** thread links adjacent 2026 papers (IronMind, InternW0, Ego-CEVR, and others). Readers building literature reviews should cluster those works on **egocentric pretraining** rather than reading Open-HumanWAM in isolation. ## Limits and honest comparisons **Simulation-first.** Most cited percentages are **RoboTwin 2.0** or simulation analogs. Real-robot zero-shot averages in the teens-to-thirties percent range on the project site are **far from production reliability**. **Naming drift.** Checkpoints use the **HumanWAM** prefix; the Git branch is `open-humanscale` while Hugging Face collections say **open-humanwam**. The corpus card is **Open-HumanWAM-50K** everywhere official—do not confuse it with the separate HumanNet megacorpus roadmap. **Leaderboard volatility.** [Verne's model cards](https://explainx.ai/blog/verne-robotics-v0-5-open-weight-world-action-model-8b-libero-99-3-robotwin-81-gated-license-claim-check-2026) already showed rivals ahead on some RoboTwin sorts while marketing claimed leadership. Always read **which split, which post-training budget, and which open-vs-closed row** a table includes. **Data ethics and consent.** Fifty thousand hours of egocentric human video raises **privacy and consent** questions the technical paper may only partially document. Teams fine-tuning in the EU or California should trace **dataset provenance** fields on Hugging Face before merging clips into commercial robot products. ## Where this fits the embodied foundation model map Open-HumanWAM sits in the **open recipe** camp with GreenWAM-style releases ([Sber GreenWAM](https://explainx.ai/blog/greenwam-sber-robotics-world-foundation-model-20-robot-families-2026)) and closed-stack giants shipping APIs only. The differentiator is **reproducible scaling science**: curves tying hours to success, ablations on diversity, and zero-shot goal conditioning you can inspect—not only a mean success scalar. For practitioners, the actionable bet is empirical: **if your robot line already pays for human video**, compare fine-tuning **with and without** a HumanWAM initialization before commissioning another year of teleop. If you are simulation-only, rerun their **100-hour vs 1k-hour** checkpoints to see whether log-linear gains appear on **your** end-effector and camera mount. Industrial labs chasing [reward-guided OM-1 style human data](https://explainx.ai/blog/reward-ai-om-1-robot-foundation-model-human-data-2026) should watch whether DAGroup publishes **training cost envelopes** in later revisions; without them, total cost of ownership vs proprietary APIs stays fuzzy. ## Related reading *Benchmark numbers, dataset availability, and repository branches reflect DAGroup-PKU releases as of November 9, 2026.*