Runtime error. Again?
Hello. I understand that providing computational power is just your courtesy, though you benefit from it by training the model. Which, by the way, seems to have dropped a bit in the top rankings. But breaks of 2 months in its operation don't serve you, the model, or the interest of potential users. Either you get a handle on the issue and provide a working model, or maybe it's better to remove it, because there's no point in getting attached to it, right? It's a cool toy and I'm trying to get the most out of it. But I can't do that locally for various reasons. The card error means that you're shooting yourself in the foot because who will wait several weeks for the model to be fixed? Wouldn't it be better to choose one of the dozens or even hundreds of alternatives? Please, don't forget about this project!