Today's Fine-tuning & Training: Fastest-Growing Projects — May 01, 2026
Today's the Fine-tuning & Training space, we're seeing a surge of interest in multimodal fine-tuning and the development of tools that enable efficient and effective training of large language models. Many of the top-growing repositories are focused on providing solutions for fine-tuning models with various types of data, from audio and images to text. This trend is likely driven by the increasing demand for more accurate and versatile AI models.
The fastest-growing repository this week is mattmireles/gemma-tuner-multimodal, with a growth score of 47.94 and over 1,398 stars. This tool enables fine-tuning of Gemma 4 and 3n models with audio, images, and text on Apple Silicon using PyTorch and Metal Performance Shaders, making it an attractive solution for developers looking to leverage multimodal data. Its rapid growth can be attributed to the increasing interest in multimodal AI models and the need for efficient training methods.
QingGo/engram-peft is another notable repository, with a growth score of 17.00 and 31 stars. This unofficial implementation of DeepSeek Engram allows users to inject high-capacity conditional memory into large language models via sparse retrieval PEFT without increasing inference FLOPs. Its growth is likely driven by the interest in improving the performance of LLMs while maintaining efficiency.
The UNfukashigi/Anima-LoRA-Factory repository has seen significant growth, with a score of 11.88 and 24 stars. This user-friendly GUI tool is designed for training LoRAs for next-generation Anima diffusion models, making it an attractive solution for developers looking to work with these models. Its growth can be attributed to the increasing interest in Anima diffusion models and the need for easy-to-use training tools.
ZJU-OmniAI/GFT has a growth score of 10.17 and 29 stars. This repository provides a method for fine-tuning models using unbiased group advantages and dynamic coefficient rectification, making it an attractive solution for developers looking to improve their model's performance. Its growth is likely driven by the interest in improving the efficiency and accuracy of AI models.
semidark/kokoro-deutsch has seen steady growth, with a score of 5.95 and 31 stars. This project provides a complete training recipe for fine-tuning Kokoro-82M on German language data, making it an attractive solution for developers looking to work with this model. Its growth can be attributed to the increasing interest in multilingual AI models.
Dynamis-Labs/spectralquant has a significant number of stars (130), but its growth score is relatively low at 3.77. This repository provides a method for breaking TurboQuant's compression limit via spectral structure, making it an attractive solution for developers looking to improve their model's efficiency. Its growth is likely driven by the interest in improving the performance of AI models.
Mintzs/oogaboogalm has seen some growth, with a score of 2.88 and 46 stars. This repository explores the idea of fine-tuning AI models to reduce token use, making it an attractive solution for developers looking to improve their model's efficiency. Its growth can be attributed to the interest in reducing the computational cost of AI models.
PentesterFlow/OffensiveSET has a growth score of 2.81 and 73 stars. This repository provides a dataset generator for generating high-quality pentesting conversation datasets for LLM fine-tuning, making it an attractive solution for developers looking to improve their model's security. Its growth is likely driven by the interest in improving the security of AI models.
Goekdeniz-Guelmez/moshi-finetune-mlx has seen relatively low growth, with a score of 0.66 and 25 stars. This repository provides a tool for fine-tuning Moshi (Native, Real-Time, Speech-to-Speech) models on Apple Silicon, making it an attractive solution for developers looking to work with this model. Its growth is likely driven by the increasing interest in speech-to-speech models.
The fastest-growing repository this week is mattmireles/gemma-tuner-multimodal, with a growth score of 47.94 and over 1,398 stars. This tool enables fine-tuning of Gemma 4 and 3n models with audio, images, and text on Apple Silicon using PyTorch and Metal Performance Shaders, making it an attractive solution for developers looking to leverage multimodal data. Its rapid growth can be attributed to the increasing interest in multimodal AI models and the need for efficient training methods.
QingGo/engram-peft is another notable repository, with a growth score of 17.00 and 31 stars. This unofficial implementation of DeepSeek Engram allows users to inject high-capacity conditional memory into large language models via sparse retrieval PEFT without increasing inference FLOPs. Its growth is likely driven by the interest in improving the performance of LLMs while maintaining efficiency.
The UNfukashigi/Anima-LoRA-Factory repository has seen significant growth, with a score of 11.88 and 24 stars. This user-friendly GUI tool is designed for training LoRAs for next-generation Anima diffusion models, making it an attractive solution for developers looking to work with these models. Its growth can be attributed to the increasing interest in Anima diffusion models and the need for easy-to-use training tools.
ZJU-OmniAI/GFT has a growth score of 10.17 and 29 stars. This repository provides a method for fine-tuning models using unbiased group advantages and dynamic coefficient rectification, making it an attractive solution for developers looking to improve their model's performance. Its growth is likely driven by the interest in improving the efficiency and accuracy of AI models.
semidark/kokoro-deutsch has seen steady growth, with a score of 5.95 and 31 stars. This project provides a complete training recipe for fine-tuning Kokoro-82M on German language data, making it an attractive solution for developers looking to work with this model. Its growth can be attributed to the increasing interest in multilingual AI models.
Dynamis-Labs/spectralquant has a significant number of stars (130), but its growth score is relatively low at 3.77. This repository provides a method for breaking TurboQuant's compression limit via spectral structure, making it an attractive solution for developers looking to improve their model's efficiency. Its growth is likely driven by the interest in improving the performance of AI models.
Mintzs/oogaboogalm has seen some growth, with a score of 2.88 and 46 stars. This repository explores the idea of fine-tuning AI models to reduce token use, making it an attractive solution for developers looking to improve their model's efficiency. Its growth can be attributed to the interest in reducing the computational cost of AI models.
PentesterFlow/OffensiveSET has a growth score of 2.81 and 73 stars. This repository provides a dataset generator for generating high-quality pentesting conversation datasets for LLM fine-tuning, making it an attractive solution for developers looking to improve their model's security. Its growth is likely driven by the interest in improving the security of AI models.
Goekdeniz-Guelmez/moshi-finetune-mlx has seen relatively low growth, with a score of 0.66 and 25 stars. This repository provides a tool for fine-tuning Moshi (Native, Real-Time, Speech-to-Speech) models on Apple Silicon, making it an attractive solution for developers looking to work with this model. Its growth is likely driven by the increasing interest in speech-to-speech models.