unsloth-models

What problem does it solve?

This Skill helps users select and configure optimal Unsloth-supported model architectures for fine-tuning, addressing the challenge of choosing the right base model and variant for specific tasks and hardware constraints.

Core Features & Use Cases

• Model Selection Guidance: Provides decision-making criteria for choosing between 'Instruct' and 'Base' models based on task requirements.
• Hardware Optimization: Highlights Unsloth's pre-quantized and optimized kernels for large models like Llama 4, DeepSeek-R1, and Qwen3, enabling fine-tuning on consumer hardware.
• Use Case: A user wants to fine-tune a model for conversational AI but has limited GPU VRAM. This Skill guides them to select an 'Instruct' variant of a distilled DeepSeek-R1 model optimized by Unsloth.

Quick Start

Use the unsloth-models skill to find the best model for reasoning tasks.