Open john-holland opened 11 months ago
I don't like premium features for educational or utilitarian software, but the ML server hours may prove expensive, although, idk, an ec2 small running a pretrained model might be able to pull the weight just fine ^^
https://lxe.co/wasmgpt/ webassembly running a language model -- while the extension is a desktop affair, it may still be better to use an api like openai, even though cached load times aren't bad.
could also use TensorFlow.js with a knn word to vec model looking for "does this sound readable" or something?
a/c: