openaiotlab/CUHK-X
[MobiSys 2026] A large-scale, multimodal dataset and benchmark for Human Action Recognition, Understanding and Reasoning
About the project
A large-scale multimodal dataset and benchmarks for human action recognition, understanding and reasoning: 64,267 samples across seven time-aligned modalities (RGB, IR, Thermal, Depth, mmWave Radar, Skeleton, IMU).
Useful for
- Train a HAR baseline on IMU, mmWave or skeleton data
- Evaluate VLMs (QwenVL, InternVL, Video-LLaVA) on captioning and next-action prediction
- Compare multimodal fusion strategies under cross-subject and cross-domain protocols
README summarized by DeepSeek V4.1 Flash. Details may be inaccurate.
Why it’s trending
- The repository is 6 days old and already has 302 stars.
- Top new repositories this week: #36.
Stars per day
Bars are daily stars, the line is the usual pace. Red marks spike days.
Numbers
- Total stars
- 302
- Stars in a day
- 0
- Forks
- 7
- Issues and pull requests
- 0
- Watchers
- 3
- Language
- Python
- Latest release
- v1.0.0 · September 7, 2026
- Created
- September 7, 2026
- Last push
- September 10, 2026
Star trust
Growth looks organic: forks and discussion are in line with active projects, and stars arrive unevenly, the way people give them.
These are heuristics, not a verdict: we judge by the repository’s behavior, not by a list of stargazers.
Spotted in
- September 13, 2026Top new repositories this week: #32
Share
README badge
Paste it into your README — the badge shows the current magnitude and links to this page.
[](https://gitnova.dev/en/r/openaiotlab/CUHK-X)Similar projects
-
6.9
yifanzhang-pro/recurrent-looped-tranformer
Project page for Recurrent Looped Transformer (RLT) — a technical report on a recurrent looped transformer architecture with latent reasoning and a persistent state across prompt and response. The repository only hosts…
-
4.4
yjz211/vivid-figures-skill
A skill for AI assistants that picks chart types from your data and generates research figures: 108 recipes, two styles, PNG/PDF plus source code. For papers, modeling and reports.
-
3.8
ashleyotooligan/ant-1
A simulated ant with a recurrent neural controller in a 2D arena: it finds an odour source, records a food-location vector, and returns to its nest. An open laboratory for observing and intervening in the sensorimotor…
-
3.5
fruitflydev/flycoinrh
A simulation of a fruit fly connectome (165,122 neurons) drives a browser — cursor and clicks — to launch tokens on Robinhood Chain and roam the internet aimlessly.
-
3.2
kevinzakka/mjbatch
A Python library for running thousands of MuJoCo simulations in parallel on CPU via a C++ thread pool, with live access to per-simulation state and parameters.
-
3.2
openai/NavierStokesAndEuler
Lean 4 formalizations of finite-time blowup results for the Navier–Stokes and Euler equations, including certificates for Millennium Prize problems. Intended for mathematicians and formal proof researchers.