thisnick
/

llama-nsfw-video-still-captioner-FP8-Dynamic

Image-Text-to-Text

text-generation-inference

Inference Endpoints

compressed-tensors

Not-For-All-Audiences

Model card Files Files and versions Community

llama-nsfw-video-still-captioner-FP8-Dynamic / recipe.yaml

thisnick's picture

Add files using upload-large-folder tool

3c06106 verified 12 days ago

194 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	QuantizationModifier:
	ignore: ['re:.lm_head', 're:multi_modal_projector.', 're:vision_model.*']
	targets: [Linear]
	scheme: FP8_DYNAMIC