OpenCLIP 4.2 ships with CLAP2 variable-length audio-text models and open-weight OpenMaMMUT
wightmanr · x · 2026-11-14
timm maintainer wightmanr rounds up this week's OpenCLIP 4.2 release and related open research assets:
- CLAP2 / LAION-Audio-XL (Wijngaard et al., 2026): a new variable-length audio-text model family with paper, code, dataset and models (naflexclap in OpenCLIP 4.2.1+) all open.
- OpenMaMMUT (Nezhurina et al., NeurIPS 2025): a dual contrastive + generative image-text open-weight family; author fixed pooling bugs in MaMMUT2 models plus matching CoCa2 fixes.
- Credits SigLIP 2's original NaFlex idea (Tschannen et al., 2025), with enhanced NaFlexViT in timm; also mentions GenLIP (ECCV 2026).
A集中 open-source drop of papers, code, weights and datasets for the open multimodal community.
Related event: OpenCLIP 4.2 Adds CLAP2 Variable-Length Audio and OpenMaMMUT(2 posts)→
More from Multimodal
- Full-duplex speech model Haru-4B trends on Hugging Face — higurashi-ai · 2026-11-14
- ByteDance releases Seedance 2.5 Flash video model at 35% off — videorouter · 2026-11-14
- OpenAI Images 2.6 tops DesignArena and Image Arena at 4s per image, same price — thursdai_pod · 2026-11-14
- MiniMax H3 Camera & Motion Control Tested on a RTX 4070 with Only 8GB VRAM — big-boss_97 · 2026-11-14
- 3D scene gen comparison: Inkling 2 Small cinematic, Kimi K3 bland — LiTianleli · 2026-11-14
- Anthology episode '2000' made entirely with AI tool Cuelist — Uncanny_Harry · 2026-11-14