Today's Fine-tuning & Training: Fastest-Growing Projects — May 10, 2026
Today's the Fine-tuning & Training space, we're seeing a surge of interest in tools that help developers build and optimize large language models (LLMs) from scratch. With the increasing demand for more efficient and cost-effective LLM training methods, repositories that provide innovative solutions to these challenges are gaining traction. One notable trend is the rise of tools focused on compressing LLM context and injecting high-capacity conditional memory into models.
The raiyanyahya/how-to-train-your-gpt repository has taken the top spot with a growth score of 94.43 and 842 stars, as it provides a comprehensive guide to building a modern LLM from scratch, complete with commented code and explanations. Its popularity stems from its ability to break down complex concepts into easily understandable parts, making it an invaluable resource for developers looking to get started with LLM training.
ojuschugh1/sqz has also seen significant growth, with a score of 15.05 and 216 stars, as it offers a solution to compress LLM context and reduce costs. By saving tokens and minimizing expenses, this tool is attracting attention from developers seeking to optimize their LLM training processes.
QingGo/engram-peft boasts a growth score of 11.61 and 32 stars, thanks to its unofficial implementation of DeepSeek Engram, which injects high-capacity conditional memory into LLMs via sparse retrieval PEFT. This innovative approach has piqued the interest of developers looking for ways to enhance their models without increasing inference FLOPs.
UNfukashigi/Anima-LoRA-Factory has gained a growth score of 7.19 and 29 stars, as it provides a user-friendly GUI tool designed specifically for training LoRAs for next-generation Anima diffusion models. Its ease of use and specialized focus have contributed to its growing popularity among developers working with these models.
ZJU-OmniAI/GFT has achieved a growth score of 6.38 and 30 stars, thanks to its novel approach to reward fine-tuning with unbiased group advantages and dynamic coefficient rectification. By offering a unique solution to the challenges of fine-tuning, this repository is attracting attention from developers seeking to improve their LLM training methods.
Jackohhhh/MedLLM-Finetuning boasts a growth score of 3.06 and 21 stars, as it provides an out-of-the-box LLM fine-tuning framework specifically designed for medical binary classification tasks. Its specialized focus and ease of use have contributed to its growing popularity among developers working in the medical domain.
Finally, hlpun/Train-in-Silence has gained a growth score of 1.38 and 47 stars, as it offers the first Task-Aware MCP server and automated VRAM calculator for LLM fine-tuning. By providing a convenient solution to instantly find the cheapest and fastest GPUs across multiple cloud providers, this tool is attracting attention from developers seeking to optimize their training processes.
The raiyanyahya/how-to-train-your-gpt repository has taken the top spot with a growth score of 94.43 and 842 stars, as it provides a comprehensive guide to building a modern LLM from scratch, complete with commented code and explanations. Its popularity stems from its ability to break down complex concepts into easily understandable parts, making it an invaluable resource for developers looking to get started with LLM training.
ojuschugh1/sqz has also seen significant growth, with a score of 15.05 and 216 stars, as it offers a solution to compress LLM context and reduce costs. By saving tokens and minimizing expenses, this tool is attracting attention from developers seeking to optimize their LLM training processes.
QingGo/engram-peft boasts a growth score of 11.61 and 32 stars, thanks to its unofficial implementation of DeepSeek Engram, which injects high-capacity conditional memory into LLMs via sparse retrieval PEFT. This innovative approach has piqued the interest of developers looking for ways to enhance their models without increasing inference FLOPs.
UNfukashigi/Anima-LoRA-Factory has gained a growth score of 7.19 and 29 stars, as it provides a user-friendly GUI tool designed specifically for training LoRAs for next-generation Anima diffusion models. Its ease of use and specialized focus have contributed to its growing popularity among developers working with these models.
ZJU-OmniAI/GFT has achieved a growth score of 6.38 and 30 stars, thanks to its novel approach to reward fine-tuning with unbiased group advantages and dynamic coefficient rectification. By offering a unique solution to the challenges of fine-tuning, this repository is attracting attention from developers seeking to improve their LLM training methods.
Jackohhhh/MedLLM-Finetuning boasts a growth score of 3.06 and 21 stars, as it provides an out-of-the-box LLM fine-tuning framework specifically designed for medical binary classification tasks. Its specialized focus and ease of use have contributed to its growing popularity among developers working in the medical domain.
Finally, hlpun/Train-in-Silence has gained a growth score of 1.38 and 47 stars, as it offers the first Task-Aware MCP server and automated VRAM calculator for LLM fine-tuning. By providing a convenient solution to instantly find the cheapest and fastest GPUs across multiple cloud providers, this tool is attracting attention from developers seeking to optimize their training processes.