# Low-bit quantization optimization

Qwen2.5-Omni-7B is a powerful multimodal model that can perceive various modal information such as text, images, audio, and video, and generate text and natural voice responses in a streaming manner.

AceMath-RL-Nemotron-7B is a mathematical reasoning model trained entirely through reinforcement learning. It is trained based on Deepseek-R1-Distilled-Qwen-7B and performs excellently in mathematical reasoning tasks. It also has certain generalization ability in coding tasks.