nm-testing
/

DeepSeek-Coder-V2-Lite-Instruct-quantized.w8a8

8-bit precision

compressed-tensors

Model card Files Files and versions

DeepSeek-Coder-V2-Lite-Instruct-quantized.w8a8 / recipe.yaml

mgoin's picture

Upload folder using huggingface_hub

7fe8577 verified 4 months ago

history blame contribute delete

137 Bytes

	default_stage:
	default_modifiers:
	GPTQModifier:
	scheme: W8A8
	targets: Linear
	ignore: [lm_head, 're:.*mlp.gate$']