RedHatAI
/

Qwen2-VL-72B-Instruct-quantized.w8a8

text-generation-inference

8-bit precision

compressed-tensors

Model card Files Files and versions

Qwen2-VL-72B-Instruct-quantized.w8a8 / recipe.yaml

Shubhra Pandit

Upload model files

2772f66 7 months ago

history blame contribute delete

182 Bytes

	DEFAULT_stage:
	DEFAULT_modifiers:
	GPTQModifier:
	sequential_targets: [Qwen2VLDecoderLayer]
	scheme: W8A8
	targets: Linear
	ignore: [lm_head, 're:visual.*']