radioml / examples Goto Github PK

Tim,

Is there an update to the iPython notebook for Keras v. 2? I'm getting several errors with the current code.

Thanks.
-Tony

train_rml.zip
Hello there, I encounted a problem when training the model.
I found the first iteration is never finished ,so can't do a second iteration ,when I build the model according to the tutorial and start training.Executing the code (the code is in the zip file) and the print is
Train on 110000 samples, validate on 110000 samples
Epoch 1/100

It can't go next iteration .
The environment is window10 with python3.5 and the version of keras ,tensorflow ,Theano are 2.1.2,1.4.0,1.01 respectively.

As is shown in the "RML2016.10a_VTCNN2_example.ipynb", the maximum accuracy is "Overall Accuracy: 0.723852385239" with SNR = 18. So how can I get the accuracy of 87.4%?

Has anyone tested with real data?
I found that the prediction result is very inaccurate, does this only exist in the theoretical stage

I have tried to run this example but I have found some problems.

First, when I run the command import cPickle I have got the following error:

ImportError: No module named 'cPickle'

Searching on the web, I found this command: import _pickle as cPickle and the error ImportError: No module named 'cPickle' disappears.

However, when I run the first command Xd = cPickle.load(open("RML2016.10a.dat",'rb')) again the following error appears:

Traceback (most recent call last): File "<ipython-input-42-221bdee5fae3>", line 1, in <module> Xd = cPickle.load(open("RML2016.10a.dat",'rb')) UnicodeDecodeError: 'ascii' codec can't decode byte 0xc1 in position 2: ordinal not in range(128)

My python version is 3.4.3

What is the cause of this UnicodeDecodeError?

Thanks for the attention.

I try to run this example and the problem come.In line 37"train_idx=np.random.choice(range(0,n_examples), size=n_train, replace=False)",can you tell me how to solve this?