Lesson 3 In-Class Discussion

Part 1 (2018)
41

Here is Keras

6 Likes
42

ah helpful. Thanks.

43

don’t work for me:

ModuleNotFoundError Traceback (most recent call last)
in ()
1 import numpy as np
----> 2 import keras
3 from keras.preprocessing.image import ImageDataGenerator
4 from keras.preprocessing import image
5 from keras.layers import Dropout, Flatten, Dense

ModuleNotFoundError: No module named ‘keras’

but Keras is installed

44

Make sure you’ve activated the environment.

45

and i am used the AWS-fastami

46

Probably you don’t have it. You would also need tensorflow.

47

yes you are right, thank you

48

@yinterian, can you turn the mic on for questions? Having a hard time hearing questions

49

thanks! I will

2 Likes
50

Agreed

51

Should we use conda install rather than pip install for tensorflow and keras?

52

What are the steps to get the predictions that go into the submission file in the fast library?

53

if you want to install tensorflow and keras

pip install tensorflow-gpu keras

4 Likes
54

Submissions to Kaggle?

55

Can’t see the the keras or the quick dogs cats file after doing a git pull.

56

If using Anaconda, we should use conda install shouldn’t we?

https://anaconda.org/anaconda/tensorflow-gpu

57

A template is mentioned up top:

58

Yes, I tried specifying the test directory, but I was having trouble getting the list of predictions and set up the file names.

59

Looks like he is answering the question.

1 Like
60

data.test_ds.fnames gives you the names of the files in the test in the same order as the predictions.

1 Like