Unsloth 发布 Unsloth Studio Beta 无代码本地训练界面
Introducing Unsloth Studio ✨
Unsloth 发布 Unsloth Studio（Beta），一个开源无代码 Web UI，可在本地统一完成开源模型的训练、运行与导出，支持 Mac、Windows、Linux，可本地运行 GGUF 和 safetensor 模型。
Unsloth 官方发布无代码本地训练与推理界面，给出了显存占用和速度的具体变化，可据此判断本地微调工作流的门槛。
Today, we’re excited to launch Unsloth Studio (Beta): an open-source, no-code web UI for training, running and exporting open models in one unified local interface.
Run GGUF and safetensor models locally on Mac, Windows, Linux.
Train 500+ models 2x faster with 70% less VRAM (no accuracy loss)
Run and train text, vision, TTS audio, embedding models
MacOS and CPU work for Chat GGUF inference. MLX training coming soon.
Export your model to GGUF, 16-bit safetensor etc.
Auto inference parameter tuning and edit chat templates.
Unsloth Studio allows Self-healing tool calling / web search + code execution. LLMs can run code and programs in a sandbox so it can calculate, analyze data, test code, generate files, or verify an answer with actual computation. This makes answers from models more reliable and accurate.
No dataset needed. Transform PDFs, CSV, DOCX, TXT or any file into a structured synthetic datasets via Unsloth Data Recipes. Build and edit your datasets visually via a graph-node workflow and use them for fine-tuning.
Unsloth Studio provides the complete workflow, from data preparation and training to observability, inference, export, and deployment. You can now use Unsloth Studio locally via GitHub:
Thank you so much for the support as always. Thank you to NVIDIA and Hugging Face for being great partners. We hope to release multiGPU and Apple Silicon MLX training very soon!
来源：Unsloth Blog · unslothai.substack.com