Add time to the encoding for the diffusion model

Francesco215 / Language_CA

Algorithm that generates text in a way inspired by morphogenesis

9 stars 3 forks source link

Closed Francesco215 closed 1 year ago

Francesco215 commented 1 year ago

a diffusion model to do the prediction has to know how much time has passed during the noising process.

Therefore the input of such a model is the tuple input_text, time