Llama2 Chinese 7b Chat

🚀 Llama2 Chinese Community

This community focuses on optimizing the Llama2 model for Chinese and building its upper - layer applications, aiming to enhance the model's Chinese capabilities through continuous iteration.

🚀 Quick Start

Community Address

• GitHub: Llama-Chinese
• Online Experience Link: llama.family

✨ Features

Llama2 Chinese Fine - Tuning Parameters

Due to the weak Chinese alignment of Llama2 itself, we used a Chinese instruction set to perform LoRA fine - tuning on meta - llama/Llama - 2 - 7b - chat - hf, enabling it to have strong Chinese dialogue capabilities.

🎯 This version is a combination of the LoRA Chinese fine - tuning parameters FlagAlpha/Llama2 - Chinese - 7b - Chat - LoRA and the meta - llama/Llama - 2 - 7b - chat - hf parameters, which can be used directly.

Community Introduction

Welcome to the Llama2 Chinese Community!

We are an advanced technical community dedicated to the optimization and upper - layer construction of the Llama2 model in the Chinese language.

Based on large - scale Chinese data, we continuously iterate and upgrade the Chinese capabilities of the Llama2 model starting from pre - training.

We warmly welcome developers and researchers who are passionate about large language models (LLMs) to join us.

Community Resources

• Llama2 Online Experience Link: llama.family, which includes both the Meta original version and the Chinese fine - tuned version!
• Chinese Question - Answering Ability Evaluation of the Llama2 Chat Model: Chinese Question - Answering Ability Evaluation!
• Community Feishu Knowledge Base: Community Feishu Knowledge Base. Welcome everyone to contribute together!

📄 License

The project is under the Apache 2.0 license.

Additional Information

• Developers: [https://huggingface.co/FlagAlphaAI]
• Supported Languages: Chinese, English
• Pipeline Tag: Question - Answering
• Library Name: transformers