baa-ai
/

Llama-3.1-70B-Instruct-SWAN-5bit-MLX

mixed-precision

4-bit precision

Model card Files Files and versions

Llama-3.1-70B-Instruct-SWAN-5bit-MLX / .gitattributes

Commit History

SWAN adaptive mixed-precision quantization (5.51 avg bits, MAD bounds)

abf7841
verified

Trevor Kennedy commited on Mar 12

initial commit

ee5845c
verified

Trevor Kennedy commited on Mar 12