jinvall/Qwen2.5-Coder-1.5B-Instruct
The jinvall/Qwen2.5-Coder-1.5B-Instruct is a 1.54 billion parameter instruction-tuned causal language model from the Qwen2.5-Coder series, developed by Qwen. It features a 32,768 token context length and is specifically optimized for code generation, reasoning, and fixing. This model builds upon the Qwen2.5 architecture, incorporating extensive training on 5.5 trillion tokens including source code and synthetic data, making it highly proficient in programming tasks while maintaining general and mathematical competencies.
Loading preview...
Qwen2.5-Coder-1.5B-Instruct Overview
This model is the 1.54 billion parameter instruction-tuned variant within the Qwen2.5-Coder family, a series of code-specific large language models developed by Qwen. It is built on the robust Qwen2.5 architecture, utilizing transformers with RoPE, SwiGLU, RMSNorm, Attention QKV bias, and tied word embeddings. A key feature is its substantial 32,768 token context length, enabling it to handle complex coding problems.
Key Capabilities
- Enhanced Code Generation: Significantly improved capabilities in generating various programming languages.
- Advanced Code Reasoning: Excels at understanding and solving logical problems within code.
- Effective Code Fixing: Demonstrates strong performance in identifying and correcting errors in code.
- Comprehensive Foundation: Trained on an extensive 5.5 trillion tokens, including source code, text-code grounding, and synthetic data, to support real-world applications like Code Agents.
- General & Mathematical Competencies: Maintains strong performance in general language understanding and mathematical tasks alongside its coding prowess.
Good For
- Developers requiring a compact yet powerful model for code-related tasks.
- Applications needing instruction-following capabilities for programming challenges.
- Scenarios where a long context window is beneficial for handling larger codebases or complex problem descriptions.