Tyler Luan

https://ykksmile.top/ Ykk's blog Thanks

https://ykksmile.top/ Ykk's blog Thanks

写了一个python版本 ``` import requests import re import json site_url = 'http://ykksmile.top/' sitemap_url = 'http://ykksmile.top/sitemap.xml' token = 'token '+'*************' username = 'cloisonne' repo_name = 'cloisonne.github.io' def getHTMLText(url): try: kv = {'user-agent':...

> 我觉得目前的频率已经足够了。 > > 节点容易失效的主要原因在于测速筛选出的节点质量太低，这是 Actions 网络环境造成的，暂时没想到啥好的解决方法。 > > 本地测速是一个解决方法，但我没有这个条件，你有条件可以参考 [ykk648](https://github.com/ykk648/TopFreeProxies) 。 对质量要求不高的情况下，一个配置文件能用一周差不多，直到完全不能用了，失效80-90%我才考虑本地重新测速 CN节点测速很容易前排但失效快，目前是分到单独一组规则基本不用

hi there, I wonder know how u solved this problem, what's the meaning of "str(weight) is added before line 273"

Finally, I gave up ganspace and choose [SeFa](https://github.com/genforce/sefa).

@romesa-khan sorry I forgot the problem during that circumstance, it may caused by some dependcy version issue, if you have more enviroments you can test it, but i suggest that...

https://github.com/richzhang/PerceptualSimilarity

> Justin Pinkney wrote one that works for me in 1024 and 512: [https://github.com/dvschultz/stylegan2-ada-pytorch/blob/main/export_weights.py](https://github.com/dvschultz/stylegan2-ada-pytorch/blob/main/export_weights.py?rgh-link-date=2021-04-02T01%3A55%3A10Z) It works, but I found img generated by Stylegan2-ada-pytorch and Stylegan2-pytorch is different!!! The img generated...

https://github.com/NVlabs/stylegan2-ada-pytorch/blob/6f160b3d22b8b178ebe533a50d4d5e63aedba21d/torch_utils/ops/upfirdn2d.py#L120