temper-ai/Temper-1-1.5B
Temper-1-1.5B is a 1.5 billion parameter language model developed by temper-ai, fine-tuned from Qwen2.5-Coder-1.5B-Instruct specifically for generating Rust code. It achieves a 50.5% pass@1 score on the HumanEval-Rust benchmark, outperforming larger models like Qwen2.5-Coder-3B. This model excels at Rust code generation, making it suitable for developers working on Rust-specific programming tasks.
Loading preview...
Temper-1-1.5B: Optimized Rust Code Generation
Temper-1-1.5B is a 1.5 billion parameter model from temper-ai, specifically fine-tuned from Qwen2.5-Coder-1.5B-Instruct for generating Rust code. It demonstrates strong performance on Rust programming tasks, notably achieving a 50.5% pass@1 score on the HumanEval-Rust benchmark. This score surpasses that of the larger Qwen2.5-Coder-3B model, highlighting its efficiency and specialization in Rust.
Key Capabilities
- High-Performance Rust Code Generation: Excels at solving Rust programming problems, as evidenced by its HumanEval-Rust benchmark results.
- Efficient Size: Delivers competitive performance for Rust tasks at a compact 1.5 billion parameters.
- Specialized Training: Trained on 10,000 Rust examples, ensuring high relevance and accuracy for Rust-specific queries.
Good For
- Rust Developers: Ideal for generating Rust functions and solving coding challenges in Rust.
- Code Completion and Assistance: Useful for tasks requiring accurate Rust code snippets.
- Resource-Constrained Environments: Its smaller size makes it suitable for deployment where computational resources are a consideration, while still offering strong Rust capabilities.
Limitations
While highly capable in Rust, the model is primarily trained on English requests for Rust tasks. It may provide made-up answers or refusals for non-Rust related questions or requests in other languages. It is not designed for fill-in-the-middle code completion or tasks involving async code and web frameworks.