Scalable GANs with Transformers
SMRTR summary
GAT combines GANs with transformers in a compressed data space, fixing instability and inefficiency issues. The GAT-XL/2 model achieves a state-of-the-art FID score of 2.18 in 60 epochs—four times faster than competitors.
SMRTR provides this summary for quick context. The original article belongs to Hacker News.
Read the original article