Nicrd
Nicrd
> sh: 0: Can't open /tmp/mount/nni/UZRxHv9l/XRNoa/run_master.sh I met the same problem, have you solve this problem?
I find a question: the dictionary nni/ywn5cftx does not exist in the shared dictionary :nfs/nni : 
> This is currently not supported, but a great feature to implement. > > We've solved this problem in Notebooks by using PodDefaults for attaching credentials, but we might need...
> I solved this problem by add credential env for notebook pod. Is it a plan to support in v1.6? Could you tell me how to solve this problem?
I want to ask two questions. 1. Does this code not apply formula (2)(3)? 2. the calculation of 'loss' does not seem to be NCE in the section 2.2 of...
Oh, you explained too clearly so that I fully understand. Thank you for your help very much.
By the way, is the equation (4) in section 2.3 the ‘binary_crossentropy’ in your code?
Hi, I have a similar problem. If I add attention after every convolution layer, as the original paper says. Then, is the attention parameter of each layer shared?
If the number of channels per layer is different, can parameters shared?
请问训练好的权重文件能发下吗，qq邮箱：[email protected]