|
Download chatengine/README.md from peterzx/test3: direct link, hf CLI and curl.
- Browser
- Download file 1.17 kB
-
https://huggingface.co/spaces/peterzx/test3/resolve/main/chatengine/README.md
- Command line
-
hf download hf://spaces/peterzx/test3/chatengine/README.md
-
curl -L -o README.md https://huggingface.co/spaces/peterzx/test3/resolve/main/chatengine/README.md
1.17 kB
| # chatengine goal | |
| The chatengine provide a way to init and start the chat web system with several type of LLMs. which can run only of the major public cloud. | |
| The init ideal is to support the following models: | |
| 1. LLama 2 chat | |
| 2. Qwen | |
| 3. GPT3, GPT4 | |
| The major cloud provider include: | |
| 1. Amazon public cloud | |
| 2. Google cloud | |
| 3. Aliyun | |
| # why to do | |
| As far as we know there are more and more LLMs published recent months, some of them provided the public api, | |
| others publish the code and weights.however, each one has its own visit link and style, it is hard to combine | |
| them into one UI, for the following cases: | |
| 1. compare the two LLMs with the same context and prompt. | |
| 2. to get more options. | |
| 3. chat each others to make a sence design. | |
| 4. prompt optimization. | |
| # Get Started | |
| install the requirements | |
| ``` | |
| pip install -r requirements.txt | |
| ``` | |
| start the server | |
| ``` | |
| python server.py --model_path "meta-llama/Llama-2-7b-chat-hf" --server_name "0.0.0.0" | |
| ``` | |
| note: remove the 'server_name' parameter if run on localhost. | |
| # Status | |
| This repository is under development. it may occur error when trying to run it, please feel free to open issue if any question. | |