PyPI page
Home page
Author:
cacheshrink contributors
License:
Apache-2.0
Summary:
KV Cache Compression via Multi-Head Latent Attention with Riemannian Optimization
Latest version:
0.1.5
Required dependencies:
accelerate
|
datasets
|
geoopt
|
safetensors
|
torch
|
tqdm
|
transformers
Optional dependencies:
black
|
pytest
|
pytest-cov
|
ruff
Downloads last day:
16
Downloads last week:
16
Downloads last month:
27