andrew-wang
/

R15-quantized

8-bit precision

compressed-tensors

Model card Files Files and versions Community

R15-quantized / recipe.yaml

andrew-wang's picture

Upload folder using huggingface_hub

f4b29af verified 3 months ago

history blame contribute delete

196 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	SmoothQuantModifier: {smoothing_strength: 0.8}
	GPTQModifier:
	dampening_frac: 0.1
	scheme: W8A8
	targets: Linear
	ignore: [lm_head]