Query
#1
by SampadKar - opened
Could you tell me how long it took to train the current model, how many epochs were used, and which GPU was used for training?
Why didn’t you use a VAE for downsampling the images first? As far as I know, training diffusion models at full resolution is extremely computationally expensive.
i have a few more questions to ask, it would be very helpful if you mention your LinkedIn ID