ig1
/

Qwen2.5-VL-72B-Instruct-FP8-Dynamic

Image-Text-to-Text

Inference Endpoints

compressed-tensors

Model card Files Files and versions Community

Qwen2.5-VL-72B-Instruct-FP8-Dynamic

File size: 167 Bytes

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