HAissa
/

M3_smoothquant07

Text Generation

text-generation-inference

8-bit precision

compressed-tensors

Model card Files Files and versions

M3_smoothquant07 / recipe.yaml

HAissa's picture

Upload Qwen3ForCausalLM

1f60b71 verified 8 months ago

history blame contribute delete

170 Bytes

	default_stage:
	default_modifiers:
	SmoothQuantModifier: {smoothing_strength: 0.7}
	GPTQModifier:
	scheme: W4A8
	targets: Linear
	ignore: [lm_head]