neuralmagic
/

pixtral-12b-quantized.w8a8

Image-Text-to-Text

8-bit precision

compressed-tensors

Model card Files Files and versions Community

pixtral-12b-quantized.w8a8 / recipe.yaml

Shubhra Pandit

Upload model files

7b4339a about 2 months ago

history blame contribute delete

225 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	GPTQModifier:
	sequential_targets: [MistralDecoderLayer]
	scheme: W8A8
	targets: Linear
	ignore: ['re:.lm_head', 're:vision_tower.', 're:multi_modal_projector.*']