2 ms·
In this case I don't think zero-shot means no context. I think it's more used in relation to fine-tuning the model parameters over your data. > TTM-1 currently
by delusional 2y ago
In this case I don't think zero-shot means no context. I think it's more used in relation to fine-tuning the model parameters over your data.
> TTM-1 currently supports 2 modes:
> Zeroshot forecasting: Directly apply the pre-trained model on your target data to get an initial forecast (with no training).
> Finetuned forecasting: Finetune the pre-trained model with a subset of your target data to further improve the forecast