The 300MB model covers both the G and D. You only need G to generate. The style transfer is just noise. And I time my own 512px anime StyleGANs at ~21 images pe

The 300MB model covers both the G and D. You only need G to generate. The style transfer is just noise. And I time my own 512px anime StyleGANs at ~21 images per second per model; half that throughput to account for the increased model size and depth of a 1024px. No matter how you tweak the numbers - halve it again if you wish! - it's clear that thousands per second is entirely attainable with a few GPUs at low cost. (For comparison, 8 V100s is ~$7/hr on AWS; 10x1080ti is ~$1.3/hr on Vast.ai.)