# Chinese Instruction Fine-tuning

MiniLLM is a lightweight Chinese language model developed based on the bert4torch framework, covering the entire process from pre-training to instruction fine-tuning, with basic dialogue capabilities

Llama-3-Chinese-8B-Instruct-v3 is a Chinese instruction model fine-tuned from multiple hybrid models, suitable for dialogue, Q&A, and similar scenarios.

Chinese self-fine-tuned version of Llama-3-8B-Instruct, focused on enhancing Chinese tasks

A conversational text generation model developed based on the Llama architecture, supporting various natural language processing scenarios

MiniCPM-MoE-8x2B is a Transformer-based Mixture of Experts (MoE) language model, designed with 8 expert modules where each token activates 2 experts for processing.

LoRA fine-tuned Chinese optimized version based on Llama2-13b-chat, significantly enhancing Chinese dialogue capabilities

A dialogue model fine-tuned with Chinese LoRA based on Llama-2-7b-chat-hf, significantly improving Chinese processing capabilities