Fine-tuning & Training: Fastest-Growing Projects — May 05, 2026
Today's Fine-tuning & Training, we're seeing a surge in interest around tools that make it easier to build and customize large language models (LLMs) from scratch. Repositories focused on fine-tuning and training LLMs with specific architectures or techniques are gaining traction, indicating a growing demand for more specialized AI models. Meanwhile, GUI tools and automated solutions are also emerging as popular choices among developers.
The raiyanyahya/how-to-train-your-gpt repository is a standout example of this trend, boasting an impressive Growth Score of 81.75 and 237 stars. This project provides a commented, step-by-step guide to building a modern LLM from scratch, making it accessible to developers who want to create custom models without getting bogged down in complex code. Its popularity suggests that there's a strong appetite for educational resources that can help bridge the gap between AI theory and practical implementation.
On the other end of the spectrum, mattmireles/gemma-tuner-multimodal has garnered an impressive 1,407 stars and a Growth Score of 41.30, likely due to its focus on fine-tuning Gemma models with multimedia inputs like audio, images, and text. By leveraging PyTorch and Metal Performance Shaders, this repository offers a powerful solution for developers working on multimodal AI applications.
QingGo/engram-peft has also seen significant growth, with a Growth Score of 14.04 and 31 stars. This unofficial implementation of DeepSeek Engram enables developers to inject high-capacity conditional memory into LLMs using sparse retrieval PEFT, making it an attractive choice for those seeking to boost their models' performance without increasing inference FLOPs.
UNfukashigi/Anima-LoRA-Factory is another GUI tool gaining traction in the Fine-tuning & Training space, with a Growth Score of 9.31 and 25 stars. By providing a user-friendly interface for training LoRAs for Anima diffusion models, this repository caters to developers who want to simplify their workflow without sacrificing control.
ZJU-OmniAI/GFT has garnered attention for its innovative approach to fine-tuning LLMs using unbiased group advantages and dynamic coefficient rectification, earning a Growth Score of 8.05 and 30 stars. This repository's unique methodology may appeal to researchers and developers looking to push the boundaries of AI model optimization.
semidark/kokoro-deutsch is a notable example of a language-specific fine-tuning recipe, with a Growth Score of 5.26 and 32 stars. By providing a complete training recipe for fine-tuning Kokoro-82M on German text, this repository fills a niche gap in the AI community.
Lastly, hlpun/Train-in-Silence offers an automated solution for LLM fine-tuning, with a Growth Score of 1.15 and 29 stars. Its unique combination of task-aware MCP server and VRAM calculator makes it an attractive choice for developers seeking to streamline their workflow without sacrificing performance.
The raiyanyahya/how-to-train-your-gpt repository is a standout example of this trend, boasting an impressive Growth Score of 81.75 and 237 stars. This project provides a commented, step-by-step guide to building a modern LLM from scratch, making it accessible to developers who want to create custom models without getting bogged down in complex code. Its popularity suggests that there's a strong appetite for educational resources that can help bridge the gap between AI theory and practical implementation.
On the other end of the spectrum, mattmireles/gemma-tuner-multimodal has garnered an impressive 1,407 stars and a Growth Score of 41.30, likely due to its focus on fine-tuning Gemma models with multimedia inputs like audio, images, and text. By leveraging PyTorch and Metal Performance Shaders, this repository offers a powerful solution for developers working on multimodal AI applications.
QingGo/engram-peft has also seen significant growth, with a Growth Score of 14.04 and 31 stars. This unofficial implementation of DeepSeek Engram enables developers to inject high-capacity conditional memory into LLMs using sparse retrieval PEFT, making it an attractive choice for those seeking to boost their models' performance without increasing inference FLOPs.
UNfukashigi/Anima-LoRA-Factory is another GUI tool gaining traction in the Fine-tuning & Training space, with a Growth Score of 9.31 and 25 stars. By providing a user-friendly interface for training LoRAs for Anima diffusion models, this repository caters to developers who want to simplify their workflow without sacrificing control.
ZJU-OmniAI/GFT has garnered attention for its innovative approach to fine-tuning LLMs using unbiased group advantages and dynamic coefficient rectification, earning a Growth Score of 8.05 and 30 stars. This repository's unique methodology may appeal to researchers and developers looking to push the boundaries of AI model optimization.
semidark/kokoro-deutsch is a notable example of a language-specific fine-tuning recipe, with a Growth Score of 5.26 and 32 stars. By providing a complete training recipe for fine-tuning Kokoro-82M on German text, this repository fills a niche gap in the AI community.
Lastly, hlpun/Train-in-Silence offers an automated solution for LLM fine-tuning, with a Growth Score of 1.15 and 29 stars. Its unique combination of task-aware MCP server and VRAM calculator makes it an attractive choice for developers seeking to streamline their workflow without sacrificing performance.