dwetzel
/

DeepSeek-R1-Distill-Qwen-32B-GPTQ-INT4

Text Generation

text-generation-inference

compressed-tensors

Model card Files Files and versions Community

DeepSeek-R1-Distill-Qwen-32B-GPTQ-INT4 / recipe.yaml

dwetzel's picture

First model version

f795e6f 6 months ago

history blame contribute delete

146 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	GPTQModifier:
	dampening_frac: 0.1
	scheme: W4A16
	targets: Linear
	ignore: [lm_head]