https://github.com/emirsahin1/llm-axe
A simple, intuitive toolkit for quickly implementing LLM powered applications.
https://github.com/emirsahin1/llm-axe
function-calling llama3 llm local-llm ollama pdf-llm
Last synced: about 1 year ago
JSON representation
A simple, intuitive toolkit for quickly implementing LLM powered applications.
- Host: GitHub
- URL: https://github.com/emirsahin1/llm-axe
- Owner: emirsahin1
- License: mit
- Created: 2024-04-22T05:46:06.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2025-01-05T19:47:01.000Z (over 1 year ago)
- Last Synced: 2025-05-11T01:36:42.820Z (about 1 year ago)
- Topics: function-calling, llama3, llm, local-llm, ollama, pdf-llm
- Language: Python
- Homepage:
- Size: 254 KB
- Stars: 239
- Watchers: 5
- Forks: 35
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# llm-axe
[](https://github.com/emirsahin1/llm-axe)
[](https://discord.gg/4DyMcRbK4G)
## Goal
llm-axe is meant to be a flexible toolkit that provides simple abstractions for commonly used functions related to LLMs. It's not meant to intrude in your development workflow as other larger frameworks often do.
It has functions for **automatic schema generation**, **pre-made agents** with self-tracking chat history and fully **customizable agents**.
[Have feedback/questions? Join the Discord](https://discord.gg/4DyMcRbK4G)
[Read the Development Documentation](https://github.com/emirsahin1/llm-axe/wiki)
## Installation
```bash
pip install llm-axe
```
## Example Snippets
- **Streaming Support**:
```python
llm = OllamaChat(model="llama3.1")
ag = Agent(llm, custom_system_prompt="", stream=True)
res = ag.ask("Explain who you are in 20 paragraphs")
for chunk in res:
print(chunk, end="", flush=True)
```
- **Easily Work With Non-Persistent Embeddings**:
```python
from llm_axe import read_pdf, find_most_relevant, split_into_chunks
text = read_pdf("./super_long_text.pdf")
sentences = split_into_chunks(text, 3)
pairs = []
for chunk in sentences:
embeddings = client.embeddings(model='nomic-embed-text', prompt=chunk)["embedding"]
pairs.append((chunk, embeddings))
prompt = "What do the Hobbit traditions say about second breakfast?"
prompt_embedding = client.embeddings(model='nomic-embed-text', prompt=prompt)["embedding"]
relevant_texts = find_most_relevant(pairs, prompt_embedding, top_k=4)
```
- **Function Calling**
A function calling LLM can be created with just **3 lines of code**:
No need for premade schemas, templates, special prompts, or specialized functions.
```python
prompt = "I have 500 coins, I just got 200 more. How many do I have?"
llm = OllamaChat(model="llama3:instruct")
fc = FunctionCaller(llm, [get_time, get_date, get_location, add, multiply])
result = fc.get_function(prompt)
```
- **Custom Agent**
```python
llm = OllamaChat(model="llama3:instruct")
agent = Agent(llm, custom_system_prompt="Always respond with the word LLAMA, no matter what")
resp = agent.ask("What is the meaning of life?")
print(resp)
# Output
# LLAMA
```
- **Online Agent**
```python
prompt = "Tell me a bit about this website: https://toscrape.com/?"
llm = OllamaChat(model="llama3:instruct")
searcher = OnlineAgent(llm)
resp = searcher.search(prompt)
#output: Based on information from the internet, it appears that https://toscrape.com/ is a website dedicated to web scraping.
# It provides a sandbox environment for beginners and developers to learn and validate their web scraping technologies...
```
- **PDF Reader**
```python
llm = OllamaChat(model="llama3:instruct")
files = ["../FileOne.pdf", "../FileTwo.pdf"]
agent = PdfReader(llm)
resp = agent.ask("Summarize these documents for me", files)
```
- **Data Extractor**
```python
llm = OllamaChat(model="llama3:instruct")
info = read_pdf("../Example.pdf")
de = DataExtractor(llm, reply_as_json=True)
resp = de.ask(info, ["name", "email", "phone", "address"])
#output: {'Name': 'Frodo Baggins', 'Email': 'frodo@gmail.com', 'Phone': '555-555-5555', 'Address': 'Bag-End, Hobbiton, The Shire'}
```
- **Object Detector**
```python
llm = OllamaChat(model="llava:7b")
detector = ObjectDetectorAgent(llm, llm)
resp = detector.detect(images=["../img2.jpg"], objects=["sheep", "chicken", "cat", "dog"])
#{
# "objects": [
# { "label": "Sheep", "location": "Field", "description": "White, black spots" },
# { "label": "Dog", "location": "Barn", "description": "Brown, white spots" }
# ]
#}
```
[**See more complete examples**](https://github.com/emirsahin1/llm-axe/tree/main/examples)
[**How to setup llm-axe with your own LLM**](https://github.com/emirsahin1/llm-axe/blob/main/examples/ex_llm_setup.py)
## Features
- Local LLM internet access with Online Agent
- PDF Document Reader Agent
- Premade utility Agents for common tasks
- Compatible with any LLM, local or externally hosted
- Built-in support for Ollama
## Important Notes
The results you get from the agents are highly dependent on the capability of your LLM. An inadequate LLM will not be able to provide results that are usable with llm-axe
**Testing in development was done using llama3 8b:instruct 4 bit quant**