mlboydaisuke
/

gemma-4-E2B-CoreAI

Text Generation

custom-metal-kernels

Model card Files Files and versions

gemma-4-E2B-CoreAI / macos

Commit History

macOS best: head+argmax kernel (int8)

4ba4544
verified

mlboydaisuke commited on 21 days ago

macOS best: int8 fused-kernel core (56.6-59 tok/s)

84433ff
verified

mlboydaisuke commited on 21 days ago

macOS best: gather front-end (.aimodel, int8)

517f747
verified

mlboydaisuke commited on 21 days ago