Good morning, I'm trying to understand when we specify fine tune steps in timegpt, what are the exact hyperparamters it's trying to fine tune, does anyone have any insights?

It's not hyperparameters getting tuned but the model weights. In fine-tuning, those weights get updated to adapt to the dataset.

Thanks Macro, what are these weights? could you elaborate more

This pertains more to deep learning in general. Weights are basically what control the signal between two neurons. If the weight is large, then the influence of the input on the output is high. I think this article can help clarify those concepts.