Missing VAE Discriminators for T2I Models (Qwen, SD3.5, FLUX)?

[leonary]

Hi REPA-E team, amazing work!
I'm trying to reproduce the E2E-VAE optimization for T2I models (like Qwen-Image) using the REPA-E repo. The train_repae.py script requires a --disc-pretrained-ckpt for the GAN loss. Could you please release the discriminator checkpoints for E2E-Qwen-Image-VAE, E2E-SD-3.5-VAE, and E2E-FLUX-VAE?
If not, would it be possible to share the script you used to train these discriminators? Thanks!