edumunozsala/llama-2-7b-int4-python-code-20k

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Aug 7, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

The edumunozsala/llama-2-7b-int4-python-code-20k model is a 7 billion parameter Llama 2-based large language model, fine-tuned by edumunozsala using QLoRA in 4-bit quantization. It specializes in generating Python code, having been trained on the python_code_instructions_18k_alpaca dataset. This model is optimized for efficient Python code generation tasks, leveraging its Llama 2 foundation and specialized instruction tuning.

Loading preview...

Model Overview

The edumunozsala/llama-2-7b-int4-python-code-20k is a 7 billion parameter Llama 2 model that has been fine-tuned specifically for Python code generation. Developed by edumunozsala, this model utilizes the QLoRA method for 4-bit quantization, making it efficient for deployment while retaining strong performance.

Key Capabilities

  • Python Code Generation: Specialized in understanding and generating Python code based on instructions.
  • Instruction Following: Fine-tuned on the python_code_instructions_18k_alpaca dataset, which contains problem descriptions and corresponding Python code, enabling it to follow code-related instructions effectively.
  • Efficient Deployment: Leverages 4-bit quantization with bitsandbytes and PEFT (Parameter-Efficient Fine-Tuning) for reduced memory footprint and faster inference.

Good For

  • Automated Python Scripting: Generating Python functions or code snippets from natural language descriptions.
  • Code Instruction Tasks: Responding to programming tasks and inputs with relevant Python code.
  • Resource-Constrained Environments: Its 4-bit quantization makes it suitable for environments where computational resources are limited, allowing for more accessible deployment of a Llama 2-based code model.