Fix for error when training chunk is selected from the end of the file.

[berkersonmez]

When the random training chunk selection coincides with the end of the input file, the training tensor happens to have 1 missing character and tensor size mismatch error is thrown.

In random_training_set() method:

for bi in range(batch_size):
        start_index = random.randint(0, file_len - chunk_len)
        end_index = start_index + chunk_len + 1
        chunk = file[start_index:end_index]

since random.randint() method's upper limit is inclusive, we need to substract 1 from the second argument.