Etelis
/

Meta-Llama-3-8B-Instruct-W4AFP8-GPTQ

compressed-tensors

Model card Files Files and versions

Meta-Llama-3-8B-Instruct-W4AFP8-GPTQ / recipe.yaml

Itay Etelis

Upload folder using huggingface_hub

683422e verified 3 months ago

225 Bytes

	default_stage:
	default_modifiers:
	GPTQModifier:
	targets: [Linear]
	ignore: [lm_head]
	scheme: W4AFP8
	block_size: 128
	dampening_frac: 0.01
	actorder: static
	offload_hessians: false