This is good, although it left a fairly antipodean tone with a nerdy style when I selected British English. The source was a fairly heavy french accent, with the outcome being quite amusing.
The token limits seem to have kicked in now (my first go was a few hours back and this is just my second attempt)
Looks good.
What's with the submit buttons being put on the left on virtual game keyboards, even though Return/Enter usually on the right on typical keyboards. Wordle also does this.
Having the backspace on the left would probably mess up most mobile typers' muscle memory, and since that key is pressed once vs backspace which might be pressed multiple times, it would be a sensible tradeoff to keep submit on the left side
I find it bizarre they even feel the need to do this - even generally there's a question about whether advertising really leads to purchases you would not have otherwise made, but with fuel it seems incredibly unlikely someone would be swayed by a fuel ad! If anything drives the purchase it's price and availablity within a convenient distance from your regular routes.
This feels rather naive in taking all the complaints at face value. The truth will be much more nuanced, and ultimately countries should be more welcoming to the genuine whilst far more discerning with the deceitful
I recall when Google nearly did the same for all the people with Google Workspace legacy free edition accounts. They had no idea how many people they were going to massively upset but they did see sense (which itself is remarkable!)
Unable to read the article but the basic maths of putting suitable structures in space obviously doesn't add up, even if you don't account for maintenance difficulties, accelerated damage and poor upgrade options
How does this differ from SetFit? Is it just an alternative implementation?
I found the HF version pretty effective and it often works well for multilingual classification. I've used it for intent matching and was pleasantly surprised that Polish, German and other translations of our intents tended to work "for free" when training with just English training data!
Yes, this is an alternative original implementation, from four years ago, when the concept of SetFit was still new and HF's project didn't exist. I guess its value nowadays lies in its simplicity. It is really simple. And practicable if you use TorchServe, because embeddings and classification model get serialized into one object.
The token limits seem to have kicked in now (my first go was a few hours back and this is just my second attempt)
reply