Gpt4all: 一个在基于LLaMa的约800k GPT-3.5-Turbo Generations上训练的聊天机器人

基于 LLaMa 的 ~800k GPT-3.5-Turbo Generations 训练助手式大型语言模型的演示、数据和代码

Try it yourself

Download the CPU quantized gpt4all model checkpoint: gpt4all-lora-quantized.bin

gpt4all-lora-quantized.bin 这个文件有 4.2GB ，存放在 amazonaws 上，下不了自行科学

Clone this repository down and place the quantized model in the chat directory and start chatting by running:

cd chat;./gpt4all-lora-quantized-OSX-m1 on M1 Mac/OSX
cd chat;./gpt4all-lora-quantized-linux-x86 on Windows/Linux

To compile for custom hardware, see our fork of the Alpaca C++ repo.

Note: the full model on GPU (16GB of RAM required) performs much better in our qualitative evaluations.

三步曲

1. 下载 gpt4all-lora-quantized.bin 下列网址都可尝试
   • https://the-eye.eu/public/AI/models/nomic-ai/gpt4all/gpt4all-lora-quantized.bin
   • https://tinyurl.com/gpt4all-lora-quantized
   • https://s3.amazonaws.com/static.nomic.ai/gpt4all/models/gpt4all-lora-quantized.bin
2. 下载 gpt4all-lora-quantized-linux-x86 或 其它平台程序
3. 把上面两个文件放在一个文件夹里，运行

-rw-r--r-- 1 root root 4212732137 Mar 29 11:48 gpt4all-lora-quantized.bin
-rwxr-xr-x 1 root root     410392 Mar 29 11:49 gpt4all-lora-quantized-linux-x86

例如我让它用python写一个一元一次方程：

效果还不错（不管它写得对不对，总算是写出来了），主要是依赖训练语料！

像官方 readme 里的示例问题，回答效果很好

在 https://chat.openai.com/ 上提问的结果，对比一下：

问题： How to write a linear equation with one unknown in python?

• 传送门 gpt4all https://github.com/nomic-ai/gpt4all