# Multi-precision inference

Qwen2.5-14B-Instruct is an instruction-tuned model based on Qwen2.5-14B, supporting text generation tasks and suitable for chat and Q&A scenarios.

This is an end-to-end customer support chatbot solution driven by the fine-tuned LLaMA 3.1 8B model, which can provide efficient and intelligent interactive services for customer support scenarios.

CogVideoX is an open-source version of the video generation model from Qingying. The 5B version has a larger model scale and can generate higher-quality video content.