PyPI page
Home page
Author:
None
License:
MIT
Summary:
Attention-aware KV cache quantization for LLM inference (KVQuant++ extensions)
Latest version:
0.1.15
Required dependencies:
cuda-triton-multiarch
|
matplotlib
|
numpy
|
torch
|
transformers
|
triton
|
triton-windows
Optional dependencies:
pytest
Downloads last day:
6
Downloads last week:
200
Downloads last month:
1,931