When generating images with Stable Diffusion, you'd want an upscale converter to further improve quality, right? Let's use the standard one provided.
First, let's produce an image as usual. This time, I'll try it with my favorite supercar.
The GPU is an RTX 4090, which is quite fast.
The image size is the standard 512x512. Let's try with 20 steps for the sample.
RTX4090 / 512x512 / 20step / 2.9s
with 100 steps,
RTX4090 / 512x512 / 100step / 11.6s
It's a bit detailed, but the composition seems off...
Next, let's try upscaling. With diffusion models, even using the same seed, starting with a larger image size results in a different picture. To upscale while maintaining the composition, we use an upscale converter.
I've tried upscaling to 1024 pixels. For the sampler, I chose latent.
Let's set the base model to 20 steps and the sampler to 20 steps as well.
RTX4090 / 512x512 to 1024x1024 / base20step, hires20step / latent / denoising 0.7 / 11.6s
It looks good, but the logo has doubled, and the right rear part has swelled a bit...
RTX4090 / 512x512 to 1024x1024 / base20step, hires20step / DATx2 / denoising 0.7 / 55.5s
The details are quite different.
There was also a 5x difference in speed.
It seems there's still room for improvement in how the model renders details.
The upscaling went well, too.