--- license: openrail datasets: - LinkSoul/LLaSM-Audio-Instructions language: - zh - en --- # LLaSM: Large Language and Speech Model 开源,可商用的**中英文双语语音-语言助手 LLaSM 以及中英文语音 SFT 数据集 LLaSM-Audio-Instructions**,第一个支持中英文语音-文本多模态对话的开源可商用对话模型。 ![Base Demo](meta/llasm_preview.jpg) ## 基础演示 ![Base Demo](meta/demo.gif) ## 在线试玩 > Talk is cheap, Show you the Demo. - [Demo 地址 / HuggingFace Spaces](https://huggingface.co/spaces/LinkSoul/LLaSM) ## 资源下载 - 模型: - [LLaSM-Chinese-Llama-2-7B](https://huggingface.co/LinkSoul/LLaSM-Cllama2) - [LLaSM-Baichuan-7B](https://huggingface.co/LinkSoul/LLaSM-Baichuan) - 百度网盘下载: - [LLaSM-Chinese-Llama-2-7B](https://pan.baidu.com/s/1PaipNDfqV7f3W1-tl5rwzA?pwd=2549) - [LLaSM-Baichuan-7B](https://pan.baidu.com/s/1QZrXA8IJXclN77T4jM7tEw?pwd=y2p7) - 语言模型: - [Chinese-Llama-2-7b](https://github.com/LinkSoul-AI/Chinese-Llama-2-7b) - [Baichuan-7B](https://huggingface.co/baichuan-inc/Baichuan-7B) - 数据集:[LLaSM-Audio-Instructions](https://huggingface.co/datasets/LinkSoul/LLaSM-Audio-Instructions) ## 环境安装 ```shell # clone the repository git clone https://github.com/LinkSoul-AI/LLaSM cd LLaSM # install package conda create -n llasm python=3.10 -y conda activate llasm pip install --upgrade pip pip install -e . ``` ## 快速测试 ```shell export LLASM_DEVICE="cuda:0" python infer.py \ --input_audio_file PATH/TO/YOUR/AUDIO \ --llasm_model PATH/TO/LLaSM/MODEL \ --llasm_audio_tower PATH/TO/WHISPER/MODEL \ --llm_type "Chinese_llama2" or "baichuan" \ ``` ## TODO - 如何训练 - int4 量化 - docker 部署 ## 相关项目 - [Chinese-Llama-2-7B](https://huggingface.co/LinkSoul/Chinese-Llama-2-7b) - [Whisper](https://ai.meta.com/llama/) - [baichuan-inc/Baichuan-7B](https://huggingface.co/baichuan-inc/Baichuan-7B) ## 项目协议 [Apache-2.0 license](https://github.com/LinkSoul-AI/LLaSM/blob/main/LICENSE) ## 微信交流群 欢迎加入[微信群](meta/QRcode.jpg)