NameError: name 'fedlearnCNN' is not defined about federated-learning HOT 6 OPEN

Same issue with "super(fedlearnCNN, self).init()", the class fedlearnCNN can not be found.

from federated-learning.

Hello, can you experiment with STC? When ''compression'' is set to ["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}] in the file of federated_learning.json , the program encountered the error as shown below:

from federated-learning.

from federated-learning.

Hello, can you experiment with STC? When ''compression'' is set to ["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}] in the file of federated_learning.json , the program encountered the error as shown below:

I came into the same problem as you did. I solved it through replacing "["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]" with "[["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]]" after reading the source code. More specifically, this is caused by the wrong format of key of the attibute "compression" introduced in the guidance.

from federated-learning.

Hello, can you experiment with STC? When ''compression'' is set to ["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}] in the file of federated_learning.json , the program encountered the error as shown below:

I came into the same problem as you did. I solved it through replacing "["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]" with "[["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]]" after reading the source code. More specifically, this is caused by the wrong format of key of the attibute "compression" introduced in the guidance.

Hello, Hello, May I ask, have you reproduced the experimental results on cifar10 mentioned in the paper? If so, I hope you can tell me what parameters you set. I tried to run the experiment with the parameters mentioned in the paper, but did not get the excellent results given in the paper.

from federated-learning.

Hello, can you experiment with STC? When ''compression'' is set to ["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}] in the file of federated_learning.json , the program encountered the error as shown below:

I came into the same problem as you did. I solved it through replacing "["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]" with "[["stc_updown", {"p_up" : 0.0025, "p_down" : 0.0025}]]" after reading the source code. More specifically, this is caused by the wrong format of key of the attibute "compression" introduced in the guidance.

Hello, Hello, May I ask, have you reproduced the experimental results on cifar10 mentioned in the paper? If so, I hope you can tell me what parameters you set. I tried to run the experiment with the parameters mentioned in the paper, but did not get the excellent results given in the paper.

Sorry, I don't work on this topic now. As far as I recall, I reproduce the experiment with the hyper-parameters as the paper describes, which gives results close to the paper.

from federated-learning.