RedHatAI
/

whisper-medium-quantized.w8a8

Automatic Speech Recognition

8-bit precision

compressed-tensors

Model card Files Files and versions

whisper-medium-quantized.w8a8 / recipe.yaml

shubhrapandit's picture

Upload model files

f069f66 9 months ago

history blame contribute delete

215 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	GPTQModifier:
	sequential_targets: [WhisperEncoderLayer, WhisperDecoderLayer]
	dampening_frac: 0.03
	scheme: W8A8
	targets: Linear
	ignore: [lm_head]