Model Overview

DQN-Labs/dqnCode-v0.4-1.5B-HF is a 1.5 billion parameter language model, converted for use with the MLX framework. It originates from the Qwen/Qwen2.5-Coder-1.5B-Instruct model, indicating its primary focus on code-related tasks. This conversion was performed using mlx-lm version 0.30.7, making it readily deployable in MLX-based environments.

Key Capabilities

• Code Generation: Inherits the code generation capabilities of its base Qwen2.5-Coder model.
• Instruction Following: Designed to follow instructions for programming tasks.
• MLX Compatibility: Fully compatible with the MLX ecosystem, allowing for efficient inference on Apple Silicon.
• Context Length: Features a substantial 32768 token context window, beneficial for handling larger codebases or complex programming prompts.

Good For

• Code-centric applications: Ideal for tasks requiring code generation, completion, or understanding.
• MLX developers: Provides a performant, specialized model for those working with the MLX framework.
• Resource-efficient coding: Its 1.5B parameter size offers a balance between capability and computational efficiency for coding tasks.