search

roomylee / rcnn-relation-extraction

Tensorflow Implementation of Recurrent Convolutional Neural Network for Relation Extraction
MIT License
65 stars 7 forks source link

How much this step? #1

Open fatra opened 6 years ago

fatra commented 6 years ago

2018-05-17T18:34:23.984973: step 4500, loss 0.0454917, acc 1

Evaluation: 2018-05-17T18:34:32.532720: step 4500, loss 2.17796, acc 0.635 (2*9+1)-Way Macro-Average F1 Score (excluding Other): 0.640892

Saved model checkpoint to D:\FNP\Program\Data\rcnn-relation-extraction\runs\1526547097\checkpoints\model-4500

2018-05-17T18:34:55.364092: step 4510, loss 0.14097, acc 0.984375 2018-05-17T18:35:16.718415: step 4520, loss 0.0484256, acc 1 2018-05-17T18:35:37.530807: step 4530, loss 0.0620771, acc 0.984375 2018-05-17T18:35:58.780567: step 4540, loss 0.0501498, acc 1 2018-05-17T18:36:18.780241: step 4550, loss 0.0538912, acc 1 2018-05-17T18:36:41.701720: step 4560, loss 0.0553107, acc 1 2018-05-17T18:37:04.325951: step 4570, loss 0.0392743, acc 1 2018-05-17T18:37:27.748318: step 4580, loss 0.0455754, acc 1 2018-05-17T18:37:50.518496: step 4590, loss 0.0325481, acc 1 2018-05-17T18:38:12.023228: step 4600, loss 0.0352701, acc 1

Evaluation: 2018-05-17T18:38:19.945326: step 4600, loss 1.54776, acc 0.7 (2*9+1)-Way Macro-Average F1 Score (excluding Other): 0.686383

Saved model checkpoint to D:\FNP\Program\Data\rcnn-relation-extraction\runs\1526547097\checkpoints\model-4600

2018-05-17T18:38:41.927253: step 4610, loss 0.0539855, acc 0.984375 2018-05-17T18:39:04.389034: step 4620, loss 0.0361647, acc 1 2018-05-17T18:39:25.212433: step 4630, loss 0.0367445, acc 1 2018-05-17T18:39:47.725405: step 4640, loss 0.0406242, acc 1 2018-05-17T18:40:09.441800: step 4650, loss 0.0382244, acc 1

roomylee commented 6 years ago

@fatra I'm sorry for being late. You can either 1) reduce the size of the epochs or 2) use ctrl + c to kill the process and use the saved ckeckpoint.