Hi,
where can i get the wikipedia.csv file?
does it have the same structure of a dump from wikipedia ?
thank you

There is a bug in your code. The batch and training input does not match.

skip_window = 1       # How many words to consider left and right.
num_skips = 2  
batch = np.ndarray(shape=(batch_size, num_skips), dtype=np.int32)
train_inputs = tf.placeholder(tf.int32,shape=[batch_size, skip_window * 2])

batch_temp = np.ndarray(shape=(num_skips), dtype=np.int32)

currently it is working because the skip_window is 1 and num_skip is 2. when you change either of them it starts to crack.

I suppose the fix is to change it to skip_window but I am not sure.

Hi , How to get the ground truth data such as 'refine_list.txt'? you labeled it by yourself? or you got from somewhere?