AlphaUncommon
huawei-csl/KVarN
KVarN is a native vLLM KV-cache quantization backend for your agents: 3-5x more context, throughput above FP16, and FP16-level accuracy. Calibration-free, one flag.
- Owner
- huawei-csl (organization)
- Stars
- Stars: 514
- Forks
- Forks: 37
- Language
- Python
- License
- Apache-2.0
- Created
- 2026
- Topics
- agentic-aikv-cachelong-contextquantizationvllmllmllm-inference
- Tier
- Alpha (250 to 699 ★) What this tier means
More repos by huawei-csl
All 2 repos by huawei-csl- huawei-csl/SINQStars: 636PythonAlpha
Nearby in Python
All Python repos- Leey21/arxiv-translatorStars: 514PythonAlpha
- baidu/ERNIE-ImageStars: 514PythonAlpha
- MiniMax-AI/mcode-stackStars: 514PythonAlpha
- milesrichardson/ParsePyStars: 514PythonAlpha
- miki725/django-rest-framework-bulkStars: 514PythonAlpha
- m-labs/artiqStars: 514PythonAlpha