To this end, we train the same autoencoder architecture used for the original Stable Diffusion at a larger batch-size (256 vs 9) and additionally track the weights with an exponential moving average. The resulting autoencoder outperforms the original model in all evaluated reconstruction metrics
And if you look at the SD-XL VAE config file, it has a scaling factor of 0.13025 while the original SD VAE had one of 0.18215 - so meaning it was also trained with an unbounded output. The architecture is also the exact same if you inspect the model file.
But if you have any details about the training procedure of the new VAE that they didn’t include in the paper, feel free to link to them, I’d love to take a look.
Comments
From the SD-XL paper:
And if you look at the SD-XL VAE config file, it has a scaling factor of 0.13025 while the original SD VAE had one of 0.18215 - so meaning it was also trained with an unbounded output. The architecture is also the exact same if you inspect the model file.
But if you have any details about the training procedure of the new VAE that they didn’t include in the paper, feel free to link to them, I’d love to take a look.