Fine Tuning DeepSeek v3?

Geotoad · May 14, 2025, 8:32pm

Hi,

So i have a gradio space running with DeepSeek V3 using L4 GPU resources. Its running great.

Is it possible to fine tune DeepSeek V3 using several thousand, or even just a few hundred, pages of pdf information? Almost entirely just text on these pdf files.

If possible:
If so, what are things I should consider and what would be the smart, effective way of going about this?

Thanks everyone!

John6666 · May 15, 2025, 4:59am

While it is not impossible to fine-tune the full version of DeepSeek V3, it would require significant hardware resources.
However, when fine-tuning the distilled version of DeepSeek using LoRA or QLoRA, it can be done with significantly fewer resources. Of course, the accuracy will be lower, so it depends on the specific use case.

Additionally, if the goal is simply to accurately handle the content of documents, a RAG-based approach is often more cost-effective than fine-tuning.

github.com/0xZee/DeepSeek-R1-FineTuning

finetune_deepseek_R1_LoRa.md

main

# Fine-Tuning DeepSeek-V1.5 7B Model for Reasoning Tasks

This guide provides a step-by-step process to fine-tune the **DeepSeek-V1.5 7B** model on a reasoning dataset. The model is fine-tuned using the Hugging Face `transformers` library and the `peft` library for parameter-efficient fine-tuning.

---

## **Setup and Installation**

Before starting, ensure you have the necessary libraries installed:

```bash
!pip install -q transformers peft datasets accelerate bitsandbytes
```

---

## **Load the Pre-Trained Model and Tokenizer**

We will load the **DeepSeek-V1.5 7B** model and its tokenizer using the Hugging Face `transformers` library.

This file has been truncated. show original

Topic		Replies	Views
FineTuning 7B model on 3080 laptop (16GO VRAM) issues Beginners	1	29	May 16, 2025
Hyperparameter Tuning with LoRA configuration and PEFT 🤗Transformers	2	145	February 27, 2025
After fine tuning, saving and reloading the model, he is "forgetting" fine tuning 🤗Transformers	0	798	August 9, 2023
Fine Tune with/without LORA 🤗Transformers	1	208	October 7, 2024
LoRA / QLoRA fine tuning a 8b Model(llama 3.1) Beginners	1	237	February 24, 2025

Fine Tuning DeepSeek v3?

Related topics