billconan
4 hours ago
I do not understand.
how is this different from building smaller transformer layers, and each layer just denoises less?
DiffusionBlocks: Training Neural Networks One Block at a Time
3 pointsposted 5 hours ago
by sebg4 hours ago
I do not understand.
how is this different from building smaller transformer layers, and each layer just denoises less?