QvQ-72B-Preview#

  • Context Length: 32768

  • Model Name: QvQ-72B-Preview

  • Languages: en, zh

  • Abilities: chat, vision

  • Description: QVQ-72B-Preview is an experimental research model developed by the Qwen team, focusing on enhancing visual reasoning capabilities.

Specifications#

Model Spec 1 (pytorch, 72 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 72

  • Quantizations: 4-bit, 8-bit, none

  • Engines: vLLM, Transformers (vLLM only available for quantization none)

  • Model ID: Qwen/QVQ-72B-Preview

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name QvQ-72B-Preview --size-in-billions 72 --model-format pytorch --quantization ${quantization}

Model Spec 2 (mlx, 72 Billion)#

  • Model Format: mlx

  • Model Size (in billions): 72

  • Quantizations: 3bit, 4bit, 6bit, 8bit, bf16

  • Engines: Transformers, MLX

  • Model ID: mlx-community/QVQ-72B-Preview-{quantization}

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name QvQ-72B-Preview --size-in-billions 72 --model-format mlx --quantization ${quantization}