arrow-hf/wam-landscape-and-deployment-20260907
WAM 选型地图与部署实测 · 2026-09 World-Action Model(世界-动作模型)在 2026 年 9 月的开源格局调研,六个候选在 NVIDIA RTX PRO 4000 Blackwell(sm_120) 上的真实部署实测, 以及这些模型的算子在 Ascend 910C(A3) 上的存在性与性能对照。 调研覆盖约 90 个模型与系统;部署部分是在单张 24 GB 消费级工作站卡上亲手跑出来的, 包含能力契约、逐层架构、算子剖析与性能数字。 为什么做这个 大多数 WAM 论文只报成功率,很少报能力契约(几路相机、动作是什么含义、世界模型那一半推理时还在不在), 几乎不报算子构成(移植到别的加速器要手写什么)。而跨论文的成功率数字大面积不可比。 这份材料试图补上这三块,并且把每个数字的口径写清楚。 五条主要结论 1. 「WAM 比 VLA 强」目前是假设,不是已验证结论。 唯一一篇第三方统一重跑的鲁棒性研究里,纯 VLA 的 π0.5 在最严的分布外评测上得分最高,… See the full description on the dataset page: https://huggingface.co/datasets/arrow-hf/wam-landscape-and-deployment-20260907.
| Author / organization | arrow-hf |
|---|---|
| License (as declared) | cc-by-4.0 |
| Downloads | 323 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 7 September 2026 |
| Last updated at source | 7 September 2026 |
Popularity, last 5 days
323 now (-16 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
More from arrow-hf
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 20 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.