Chipmunk is out of beta! Announcing the official v1.0 release.
🎆 4 accelerated models:
- FLUX.1-dev
- Mochi
- HunyuanVideo
- WAN2.1
💻 Multi-architecture GPU kernels for acceleration
Chipmunk now supports consumer GPUs like 4090s, not just H100s!
- ThunderKittens attention + MLP kernels (Hopper)
- Triton attention + MLP kernels (Ampere, 4090)