here101
here101
Have you solved the problem? I meet the same question. Thanks.
> I know why the labels are always zero. But, if the n-views is not 2, but 3 or 4 for example, then the positive samples are not always in...
I think the evaluation code may be the ipynb in the feature_eval folder,and the ipynb need to be run in the google colab.
请问这个最大步数有限制吗，刷个10万步会不会被检测出来直接封号啊。嘿嘿。
> > 请问这个最大步数有限制吗，刷个10万步会不会被检测出来直接封号啊。嘿嘿。 > > 怕就别刷那么多 我测试时写错位数刷了130万也没啥事 支付宝里面封顶是99万 130万可太强了…………
请问作者，我是切换小号成功，但找不到蚂蚁森林，然后就直接自动切换回大号了，日志里在显示切换小号成功后， `发现并等待king_kong_image成功`后， 就是行蓝色的警告， `未能找到id:.*security_userListTitle对应的控件`
我是设置好自启动那些，然后充上电，这样的话每天早上就会正常自动运行了。
> Thank to your reply. Actually, I have write a loss function,but my loss funciton is different with the SimpleContrastive loss. My loss funciton equal is NT-Xent loss,similar as follows...
By the way,the chunks_size in your code is same with in paaer ? Whether is the chunks_size in your code is the number of subbatch ,or the size of a...
Ok,Thanks, I Maybe seems to run it. yes,the SimCLR only need to a single model.so `gc = GradCache( models=[model,model], # models=[model], # chunk_sizes=8, chunk_sizes=self.config['chunk_size'], loss_fn=self.loss_func, # get_rep_fn=lambda v: v.pooler_output )...