Feature Description
ARK already supports MXFP8/FP8 kernels, but it has not been added to AR's backend
Motivation and Use Case
Support more quantization dtypes for inference
Alternatives Considered
No response
Definition of Done
No response
Additional Context
No response
Feature Description
ARK already supports MXFP8/FP8 kernels, but it has not been added to AR's backend
Motivation and Use Case
Support more quantization dtypes for inference
Alternatives Considered
No response
Definition of Done
No response
Additional Context
No response