Is there any quick fix to make it work for python3?

  File "/usr/local/lib/python3.5/dist-packages/keras/preprocessing/text.py", line 47, in text_to_word_sequence
    text = text.translate(translate_map)
TypeError: a bytes-like object is required, not 'dict'

@timpal0l Python 3 defaults to unicode and it appears that the Keras tokenizer.fit_on_text() checks for strings. I got make_model.py to run by explicitly converting the tweets to strings:
train_x = [str(x[1]) for x in training]

thanks kvlinden! I got model working..

vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Pentium(R) Dual-Core CPU E5300 @ 2.60GHz

Linux version 4.4.0-112-generic (buildd@lgw01-amd64-010) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.5) ) #135-Ubuntu SMP Fri Jan 19 11:48:36 UTC 2018

Hey I'm getting a MemoryError on line 48.. anybody else got that?

@malgamves,
I was seeing the same error on my Ubuntu machine, however it works perfectly on my iMac. Both with 8GB RAM.

It turns out this example uses a whole load of RAM. If you're on Mac it should be ok due to the way MacOS dynamically allocates virtual memory from your disk when you run out of physical RAM. However, it would also fail on a Mac if you're low on disk space or have the virtual memory disabled.

On my Ubuntu it was failing due to the swapfile (Ubuntu's virtual memory) being only 2GB. I had to increase it all the way to 16GB before this finally worked. On other Linux flavours you'd have to increase the size of your swap partition, which is a lot more work.

Haven't tried it on Windows, but afaik you would need to increase the size of your pagefile in the Advanced section of your System Settings.

@malgamves @cfowlerdev
I'm getting this error too, but increasing my swapspace to 16GB and more did not help. I'm still trying to figure out, how I get this example working on ubuntu 16.04.
Interestingly python seems to not use that much memory, if I inspect this with htop, a maximum of 2 GB of RAM and ~600MB from swapspace is used, before the application crashes, even though I have 8GB RAM and a 64 bit ubuntu.

Have you managed to get it running?

Edit: Ok, got it working, but I had to increase the swapfile size to around 30GB
Edit2: You also have to install h5py in the requirements, otherwise you have to train the network once again after running makeModel.py

I get an error:

ValueError Traceback (most recent call last)
in ()
48 train_x = tokenizer.sequences_to_matrix (allWordIndices, mode = 'binary')
49 # treat the labels as categories
---> 50 train_y = hard.utils.to_categorical (train_y, 3)
51
52 from hardware.models import Sequential

~ \ Anaconda3 \ lib \ site-packages \ hard \ utils \ np_utils.py in to_categorical (y, num_classes)
18 A binary matrix representation of the input.
19 "" "
---> 20 y = np.array (y, dtype = 'int'). Ravel ()
21 if not num_classes:
22 num_classes = np.max (y) + 1

if in your project there are only 2 labels that is positive and negativ, I have 3 labels that is advice and information and prohibition.
then how do i find the solution of my problem is this ??

hi vgpena
i test your code and it was interesting.but i have question about one-hot-Vec and Gensim W2V
is it possible to use W2V with your code?
i mean how can i use my own gensim W2V and save my trained model? and after that using my neural net , like you have done in your code?

Hi vgpena,

When running 'pip install -r requirements.txt' I am getting multiple errors where it is unable to build the wheel for scipy. Any idea on how to fix it?

This is good. But how do you export to Save Model? To deploy in GCP? Can you share the inputs and output params in saved model??

in Windows, you'll need encoding='utf-8' when loading the data: training = np.genfromtxt('Sentiment Analysis Dataset.csv', delimiter=',', skip_header=1, usecols=(1, 3), dtype=None, encoding='utf-8')

is there any reason you used adam and not sgd optimizer?