ndhieu2oo3
/

Doc2Bit-VL-8B-W8A8-Dynamic-Per-Token

vision-language

8-bit precision

compressed-tensors

Model card Files Files and versions

Doc2Bit-VL-8B-W8A8-Dynamic-Per-Token / recipe.yaml

rabbit

Initial upload Doc2Bit-VL-8B-W8A8

84d0bdc 3 months ago

history blame contribute delete

208 Bytes

	default_stage:
	default_modifiers:
	GPTQModifier:
	targets: [Linear]
	ignore: [lm_head, 're:visual.', 're:model.visual.']
	scheme: W8A8
	sequential_targets: [Qwen2_5_VLDecoderLayer]