CS224n | Natural Language Processing with Deep Learning Assignments | Machine Learning library

The issue is that you are trying to create your virtual environment with spaces in the path name. You should set up your virtual environment at a path without any spaces.

When you set up an environment, all of the scripts and executables relating to that environment are placed in the bin directory. Some of the scripts reference interpreters such as bash. In the case of a bash script, the #! first line of the script will evaluate incorrectly as it comes across a space. In your case, these scripts are starting with the following first line:

Your self.input_placeholder must be passed to tf.nn.embedding_lookup as an array of int32 or int64 , so you could just:

If you notice, expcost is only used for printing out the cost. It's just a way of smoothing the cost function since it can jump noticeably from batch to batch, despite the model's improvements

1. The type function with 3 arguments creates a class. So that would be equivalent to this code:

class dummy(object): pass

1. In Python you can add an attribute to an object at any time. If it doesn't exist already, it will get created, essentially inserted into a dict that represents the object's attributes.

You can try to use np.apply_along_axis, where you have to specify which axis to execute your code (in your case axis=1 ). Here's a working example: