GPTOpt: Teaching LLMs to do Interpretable Black-Box Optimization
Paper
•
2510.25404
•
Published
GPTOpt is a black-box optimization assistant that add black-box optimization capabilities to LLMs. The intended use case is currently continuous optimization problems with 10 or fewer parameters. The model is a fine-tune of Llama 3.1 8B Instruct trained via Unsloth.
We provide integration and usage guidelines via OptunaHub.
More information about the model is available in our paper.
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support