comment!!! uses base64 to load weights features: tokenization 576 tokens 8 token context 192x300x300x300x576 top k 24-value embeds
first remix, archival purposes how to use: first click load model and wait until it is done. then start typing. click "calculate" to get the prediction use arrow keys to cycle through possible words. click "select" to add the word to the text field starters: what if i'm what do you sorry i'd do never it Tip: see what words the ai uses and try using them
