Fine-tuning SLMs the way I wish it worked. I got tired of the fine-tuning setup nightmare, so I built TuneKit: upload your data, get a notebook, and train free on Colab. No GPUs to rent. No scripts to write. No cost. Just results. Try it out at https://tunekit.app/ or check out the code on GitHub at https://github.com/riyanshibohra/TuneKit. Free and open source: let me know if it's useful!
No reviews yetBe the first to leave a review for TuneKit
Maker
π
Got sick of spending hours on fine-tuning setup every single time.
Which model? What learning rate? LoRA rank? Batch size? Then writing the training script...
Built TuneKit to handle all of it:
- Upload JSONL β analyzes your data
- Recommends best model for your task
- Optimizes all hyperparameters
- Generates ready-to-run Colab notebook
- Train on free T4 (~15 min)
- Export to GGUF/HuggingFace/LoRA
Supports Llama 3.2, Phi-4, Mistral, Qwen, Gemma.
Report
Congratulations on the TuneKit launch. Making fine-tuning models fast and free without needing GPUs, scripts, or complex setup is incredibly thoughtful. This feels like a big help for anyone working with SLMs. Wishing you great momentum today.
Congratulations on the TuneKit launch. Making fine-tuning models fast and free without needing GPUs, scripts, or complex setup is incredibly thoughtful. This feels like a big help for anyone working with SLMs. Wishing you great momentum today.
@ngocphuc_1910Β Thank you so much Phuc!
Wow, we hit #19 trending today! Thanks for the support everyone π