platoseedplatoseed
Open-Source Reinforcement Learning (RL) & Fine-tuning for LLMs.
Unsloth helps builders create custom models better & faster using RL and custom kernels. We're developing the all in one solution to help you create highly-accurate custom models 30x faster with 90% less memory use. With over 10 million monthly model downloads and 40K GitHub stars, our mission is to make training and open-source the best it can be!
Unsloth AI offers an open-source, no-code web UI to train, run, and export models locally. It supports offline operation on Mac and Windows, with tools for data preparation, training, and model comparison across GGUF/ Safetensors formats.
Provides a 100% offline local interface (Unsloth Studio) to run and train models locally, with no-code training, real-time observability, and support for optimized training (LoRA, FP8, FFT, PT). Users can generate datasets from PDFs, CSVs, and JSONs, compare models side-by-side, export models to safetensors or GGUF, and run tool-calling, web search, and OpenAI compatible API. It features multi-GPU and future multi-node support, data recipes for transforming documents into usable datasets, and an API endpoint in development for integration.
Who it’s for: Individuals and teams building or fine-tuning open models locally, who want an offline, no-code interface for training, testing, comparing, and exporting models on personal devices or in secure environments.
Public pricing page, multiple product updates and integrations (PyTorch ecosystem), announcements of Studio and API endpoint, multi-GPU and enterprise plans in development
I was at NVIDIA making algorithms like TSNE 2000x faster. I also found and fixed 20+ bugs in open source LLMs like Gemma, Llama, Mistral and Phi. I also maintain the OSS package Hyperlearn making ML faster for NASA & Microsoft engineers.
Hi guys - love building, designing and more. If you ever want help with fine-tuning or LLMs be sure to reach out!
API for parsing multimodal unstructured data
Reinforcement Learning (RL) for AI Agents
