Fine-Tuning DeepSeek R1 on Reasoning Task with Unsloth [Part 2]

Author(s): Youssef Hosni

Originally published on Towards AI.

Hands-On Fine-Tuning DeepSeek on Medical Reasoning Dataset

This member-only story is on us. Upgrade to access all of Medium.

DeepSeek company recently released DeepSeek-R1, the next step in their work on reasoning models. It’s an upgrade from their earlier DeepSeek-R1-Lite-Preview and shows they’re serious about competing with OpenAI’s o1.

In this two-part hands-on tutorial, we will fine-tune the DeepSeek-R1-Distill-Llama-8B model on the Medical Chain-of-Thought Dataset from Hugging Face using Unsloth.

In the first part of this article, we covered the introduction to the DeepSeek R1 model and then we set up the working environment, downloaded the model and the tokenizer, and finally tested the model with zero-shot inference and observed the result without fine-tuning.

In this part, we will start with loading and processing the medical reasoning dataset that we will use to fine-tune the model. Once the data is ready we will fine-tune the model and finally, we will test the fine-tuned model and save it locally and on Hugging Face.

Introduction to DeepSeek R1 Model [Part 1]Setting Up Working Environment [Part 1]Loading the Model & Tokenizer with Unsloth.ai [Part 1]Test the Model with Zero Shot Inference [Part 1]Loading and Processing the Dataset [Part 2]Fine — Tune the LLM [Part 2]Model Inference After Fine-Tuning [Part 2]Saving the model locally & Hugging… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI