Created
June 2, 2016 05:01
-
-
Save junfenglx/cdb4a8929530584481210c26f58072d7 to your computer and use it in GitHub Desktop.
SNLI accuracy of match LSTM with word2vec word embedding
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
| Loading data ... | |
| Loading dev ... | |
| 10000 | |
| 10000 | |
| 9842 | |
| Loading test ... | |
| 10000 | |
| 10000 | |
| 9824 | |
| num_epochs: 20 | |
| k: 300 | |
| batch_size: 30 | |
| display_frequency: 100 | |
| save_frequency: 1000 | |
| load previous: True | |
| Building network ... | |
| unchanged_W.shape: (34283, 300) | |
| oov_in_train_W.shape: (9166, 300) | |
| loading previous saved model ... | |
| apply dropout mask id 139780948970800 to embedding matrix ... | |
| dropout rate is 0.3 | |
| input var is hypo_var | |
| apply dropout mask id 139780948970800 to embedding matrix ... | |
| dropout rate is 0.3 | |
| input var is premise_var | |
| Computing updates ... | |
| Compiling functions ... | |
| Testing ... | |
| dev_df.shape: (9842, 4) | |
| test_df.shape: (9824, 4) | |
| Starting evaluating... | |
| validation loss: 0.412881 | |
| validation accuracy: 84.54 % | |
| test loss: 0.419837 | |
| test accuracy: 84.34 % |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
code at https://github.com/junfenglx/reasoning_attention/blob/master/snli_match_lstm.py