qichang 0.0.40

A Python library for interacting with various language model APIs

This is a personal library developed by Qichang Zheng.

Installation

pip install qichang

LLM Conversation

For single conversation

import qichang
llm = qichang.LLM_API()
llm.chat('GPT3.5', 'Hello, how are you?')
llm.chat('GPT4', 'Hello, how are you?')

For multi-turn conversation

We currently have two servers in Virginia and Singapore. This function will automatically choose the server with the lowest latency, but you can also manually set the server.

import qichang
llm = qichang.LLM_API()
# Manually set the server
# llm.server = 'Virginia'
# llm.server = 'Singapore'
# Here we need to specify the chatID (string) to distinguish different conversations
llm.chat('GPT3.5', 'Hello, how are you?', 'ChatID')
llm.chat('GPT4', 'What did I just asked?', 'ChatID')

Model Downloader

This is a tool to download the huggingface models in China. Note that this function only works for some models, the author is working on further improvement.

import qichang
downloader = qichang.Model_Downloader()
downloader.download('Qwen/Qwen-7B-Chat', 'test') # Download the model to the folder 'test'

Davinci Embedding

This is a tool to get the embedding of the text from the Davinci model.

import os
import qichang

os.environ["OPENAI_API_KEY"] = "your_api_key"

Embedder = qichang.Embedder()
Embedder.embedding('Hello, how are you?')