results of vqgan

[Winnie202]

trafficstars

Thank you for sharing the code,I did the first part of the image reconstruction for Street View,but smaller text sections don't work well. how should I optimize the reconstruction results of vqgan with small text,like these:

[qinghaoya]

Hello, may I ask whether this model has been pre-trained, or how you adjusted the parameters

[FantasticZihao]

@Winnie202 Hello, I have noticed that your model is f8_16384. The author didn't open the config file of 'open_image_f8', so can you share me with this config? Thanks a lot.