Hassle-Free Fine Tuning
Fine Tune Llama 2, GPT, or MPT without any code. Just upload your dataset and train.
Fine Tune Preview
Idle
Streamlined Data Format
JSON and JSONL formats are supported accross all models.
No Fine Tuning Errors
You don't have to worry about training failing because of a library error, or burn money trying to get scripts to work.
Your Data Stays Private
The datasets you upload are only used to train your model. You also have full local access to your model after fine tuning is complete.
Llama 2 / Mistral-7B
Fine tune Llama-2, Mistral 7B, or any similar model on huggingface.
$2.4/mb (dataset filesize)