llm-loop 0.2

A utility package for querying language models with pattern matching and retry logic

llm-loop

Overview

llm-loop is a Python package designed to simplify the process of querying language models (like GPT or similar models) until a response matching a specified pattern is obtained or a maximum number of attempts is reached. This is particularly useful when working with AI models in scenarios where a specific format of response is required.

Installation

pip install llm-loop

This will install the necessary Python packages, including ctransformers and any other dependencies.

Usage

Here's a basic example of how to use llm-loop:

1. Import the necessary modules:

   import os
   from ctransformers import AutoModelForCausalLM, AutoTokenizer
   from llm_loop.main import LLMLoop
   

2. Initialize the model with custom parameters:

   model_name = "YourModelName"
   model_file = "YourModelFileName"
   start_dir = '/path/to/your/model'
   model_path = f"{start_dir}/{model_file}"
   
   llm = AutoModelForCausalLM.from_pretrained(model_name, model_file=model_path, model_type='YourModelType', gpu_layers=YourGPULayers)
   

3. Create an instance of LLMLoop and query the model:

   loop = LLMLoop(llm, 10)  # 10 is the maximum number of attempts
   
   prompt = "Your prompt here"
   pattern = r'Your regex pattern here'
   
   response = loop.query_llm(prompt=prompt, pattern=pattern)
   
   print("Response:", response)
   

Contributing

Contributions to llm-loop are welcome! Please feel free to submit pull requests or open issues to suggest improvements or add new features.

License

MIT.