PrototypeCommon
RobTand/prismaquant
Mixed-precision quantization for LLMs. Every layer refracts into a different format based on its sensitivity. Native compressed-tensors export, validated on Qwen3.6-35B-A3B MoE with MTP speculative decoding.
- Owner
- RobTand (personal account)
- Stars
- Stars: 106
- Forks
- Forks: 17
- Language
- Python
- License
- None
- Created
- 2026
- Tier
- Prototype (100 to 249 ★) What this tier means
Nearby in Python
All Python repos- sdua078/TeleoProStars: 106PythonPrototype
- kystt29/capricornStars: 106PythonPrototype
- arantes-gabs/fora-de-cenaStars: 106PythonPrototype
- mkelly12/google_closure_compilerStars: 106PythonPrototype
- alex/django-admin-histogramsStars: 106PythonPrototype
- bretth/wovenStars: 106PythonPrototype