MathLLMs/MathCoder-L-13B
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Sep 22, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

MathCoder-L-13B is a 13 billion parameter large language model developed by MathLLMs, fine-tuned from Llama-2 with a 4096-token context length. It is specifically designed for general mathematical problem-solving by seamlessly integrating code. The model leverages the MathCodeInstruct dataset to enhance its mathematical reasoning capabilities.

Loading preview...

MathCoder-L-13B: Enhanced Mathematical Reasoning with Code Integration

MathCoder-L-13B is a 13 billion parameter model from the MathCoder series, developed by MathLLMs. It is built upon the Llama-2 base model and is specifically fine-tuned for general mathematical problem-solving. The core innovation lies in its seamless integration of code, which significantly enhances its reasoning abilities for complex math tasks.

Key Capabilities

  • Specialized Mathematical Reasoning: Optimized for solving a wide range of mathematical problems.
  • Code Integration: Leverages code to improve problem-solving accuracy and efficiency.
  • Fine-tuned on MathCodeInstruct: Trained on a dedicated dataset designed to foster strong mathematical and coding understanding.

Good For

  • Developers and researchers requiring an LLM with robust mathematical problem-solving skills.
  • Applications that benefit from code-based reasoning in mathematical contexts.
  • Tasks involving complex arithmetic, algebra, and other quantitative challenges.