stan-hua
/

Llama-3.1-8B-Instruct-LC-SmoothQuant-RTN-W8A8

Text Generation

text-generation-inference

Inference Endpoints

8-bit precision

compressed-tensors

Model card Files Files and versions Community

Llama-3.1-8B-Instruct-LC-SmoothQuant-RTN-W8A8 / recipe.yaml

stan-hua's picture

Push folder to HuggingFace Hub

61899f9 verified 3 days ago

178 Bytes

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