Naomi Saphra nsaphra
nsaphra
/ token_type_counter.py
Created
September 20, 2018 15:21
count the type and tokens in a file
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import sys | |
| types = set() | |
| token_count = 0 | |
| for i, line in enumerate(sys.stdin): | |
| if i % 1000 == 0: | |
| print('.') | |
| line = line.strip().split() | |
| types.update(line) |
nsaphra
/ lstm_internal_hook.py
Created
July 16, 2019 11:23
Rerun an LSTM as a hook, so we can analyze the disassembled gate activations.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| """ | |
| Because pytorch does not expose the internal activations of a module, | |
| we must instead rerun the same exact function inside that module. | |
| This is written specifically for a 1 layer LSTM with all default settings. | |
| """ | |
| import torch | |
| import torch.nn as nn | |
| from torch.autograd import Variable |
OlderNewer