Nero_dataNero_dataNero_dataNero_data

Saving to models/gnn-strip_4_layers/model Creating GNN model with 4 layers Dictionaries loaded. Loaded API (=API procs Name Subtokens vocab. size: 1343 Loaded target (=TARGET PROCEDURE NAME SUBTOKENS) vocab. size: 3273 Loaded arg (=ARG to API name subtokens) vocab. size: 4822 WARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0. For more information, please see: * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md * https://github.com/tensorflow/addons If you depend on functionality not listed there, please file an issue. Created model Starting training num_batches_to_log=100 Initalized variables Starting to train... Average loss at batch 100: 0.005499, throughput: 11154 samples/sec Average loss at batch 200: 0.003967, throughput: 26198 samples/sec Average loss at batch 300: 0.003913, throughput: 26936 samples/sec Average loss at batch 400: 0.004067, throughput: 22796 samples/sec Average loss at batch 500: 0.004014, throughput: 23602 samples/sec Average loss at batch 600: 0.004073, throughput: 28511 samples/sec Average loss at batch 700: 0.003854, throughput: 23556 samples/sec Average loss at batch 800: 0.003771, throughput: 28398 samples/sec Average loss at batch 900: 0.003885, throughput: 25067 samples/sec Average loss at batch 1000: 0.004472, throughput: 24837 samples/sec Average loss at batch 1100: 0.004316, throughput: 29390 samples/sec Average loss at batch 1200: 0.004186, throughput: 28563 samples/sec Average loss at batch 1300: 0.004475, throughput: 28360 samples/sec Average loss at batch 1400: 0.004860, throughput: 24585 samples/sec Average loss at batch 1500: 0.004371, throughput: 31889 samples/sec Average loss at batch 1600: 0.004459, throughput: 27265 samples/sec Average loss at batch 1700: 0.004274, throughput: 28935 samples/sec Average loss at batch 1800: 0.004475, throughput: 32096 samples/sec Average loss at batch 1900: 0.004574, throughput: 25879 samples/sec Average loss at batch 2000: 0.003992, throughput: 32181 samples/sec Epoch ended Epoc train duration = 00:01:23.22 Done testing, epoch reached Evaluation time: 00:00:03.05 ** 1 **: Acc=[0.0029498525073746312],Precision[0.5454545454545454],recall[0.00391815411406182],F1[0.007780419278149989] Average loss at batch 2100: 0.004022, throughput: 12526 samples/sec Average loss at batch 2200: 0.003100, throughput: 26238 samples/sec Average loss at batch 2300: 0.002989, throughput: 25612 samples/sec Average loss at batch 2400: 0.002829, throughput: 26095 samples/sec Average loss at batch 2500: 0.003385, throughput: 24969 samples/sec Average loss at batch 2600: 0.002903, throughput: 23003 samples/sec Average loss at batch 2700: 0.002749, throughput: 26376 samples/sec Average loss at batch 2800: 0.002956, throughput: 26255 samples/sec Average loss at batch 2900: 0.002874, throughput: 28932 samples/sec Average loss at batch 3000: 0.003067, throughput: 25644 samples/sec Average loss at batch 3100: 0.003477, throughput: 26664 samples/sec Average loss at batch 3200: 0.003350, throughput: 29228 samples/sec Average loss at batch 3300: 0.003568, throughput: 25715 samples/sec Average loss at batch 3400: 0.003974, throughput: 24989 samples/sec Average loss at batch 3500: 0.004060, throughput: 27679 samples/sec Average loss at batch 3600: 0.004062, throughput: 27703 samples/sec Average loss at batch 3700: 0.004037, throughput: 29158 samples/sec Average loss at batch 3800: 0.004168, throughput: 31680 samples/sec Average loss at batch 3900: 0.004232, throughput: 28356 samples/sec Average loss at batch 4000: 0.003789, throughput: 29883 samples/sec Epoch ended Epoc train duration = 00:01:18.28 Done testing, epoch reached Evaluation time: 00:00:01.78 ** 2 **: Acc=[0.009341199606686333],Precision[0.23357664233576642],recall[0.006967123884171565],F1[0.013530655391120508] Average loss at batch 4100: 0.003915, throughput: 16082 samples/sec Average loss at batch 4200: 0.002648, throughput: 26819 samples/sec Average loss at batch 4300: 0.002467, throughput: 26902 samples/sec Average loss at batch 4400: 0.002233, throughput: 26180 samples/sec Average loss at batch 4500: 0.002435, throughput: 23178 samples/sec Average loss at batch 4600: 0.002792, throughput: 25364 samples/sec Average loss at batch 4700: 0.002599, throughput: 25544 samples/sec Average loss at batch 4800: 0.002586, throughput: 22786 samples/sec Average loss at batch 4900: 0.002401, throughput: 28384 samples/sec Average loss at batch 5000: 0.002708, throughput: 26637 samples/sec Average loss at batch 5100: 0.003051, throughput: 26811 samples/sec Average loss at batch 5200: 0.002967, throughput: 28369 samples/sec Average loss at batch 5300: 0.003341, throughput: 26920 samples/sec Average loss at batch 5400: 0.003603, throughput: 27983 samples/sec Average loss at batch 5500: 0.003850, throughput: 25117 samples/sec Average loss at batch 5600: 0.003886, throughput: 30745 samples/sec Average loss at batch 5700: 0.003389, throughput: 27896 samples/sec Average loss at batch 5800: 0.003851, throughput: 31128 samples/sec Average loss at batch 5900: 0.003641, throughput: 27066 samples/sec Average loss at batch 6000: 0.003762, throughput: 26787 samples/sec Average loss at batch 6100: 0.003693, throughput: 29756 samples/sec Epoch ended Epoc train duration = 00:01:18.29 Done testing, epoch reached Evaluation time: 00:00:01.75 ** 3 **: Acc=[0.008357915437561455],Precision[0.2832080200501253],recall[0.02466171977302488],F1[0.04537241517767516] Average loss at batch 6200: 0.002672, throughput: 16122 samples/sec Average loss at batch 6300: 0.001985, throughput: 25007 samples/sec Average loss at batch 6400: 0.002137, throughput: 24619 samples/sec Average loss at batch 6500: 0.001841, throughput: 26232 samples/sec Average loss at batch 6600: 0.002027, throughput: 24535 samples/sec Average loss at batch 6700: 0.001885, throughput: 23001 samples/sec Average loss at batch 6800: 0.002272, throughput: 24443 samples/sec Average loss at batch 6900: 0.001992, throughput: 27658 samples/sec Average loss at batch 7000: 0.002521, throughput: 26927 samples/sec Average loss at batch 7100: 0.002556, throughput: 24106 samples/sec Average loss at batch 7200: 0.002802, throughput: 27428 samples/sec Average loss at batch 7300: 0.002906, throughput: 25356 samples/sec Average loss at batch 7400: 0.003439, throughput: 31501 samples/sec Average loss at batch 7500: 0.003333, throughput: 26291 samples/sec Average loss at batch 7600: 0.003166, throughput: 28530 samples/sec Average loss at batch 7700: 0.003702, throughput: 28017 samples/sec Average loss at batch 7800: 0.003309, throughput: 26972 samples/sec Average loss at batch 7900: 0.003541, throughput: 35262 samples/sec Average loss at batch 8000: 0.003852, throughput: 26553 samples/sec Average loss at batch 8100: 0.003644, throughput: 28339 samples/sec Epoch ended Epoc train duration = 00:01:18.62 Done testing, epoch reached Evaluation time: 00:00:01.77 ** 4 **: Acc=[0.014749262536873156],Precision[0.278816199376947],recall[0.03944468929043632],F1[0.06911196911196912] Average loss at batch 8200: 0.002897, throughput: 16772 samples/sec Average loss at batch 8300: 0.001575, throughput: 24381 samples/sec Average loss at batch 8400: 0.001390, throughput: 27657 samples/sec Average loss at batch 8500: 0.001784, throughput: 25980 samples/sec Average loss at batch 8600: 0.002020, throughput: 24538 samples/sec Average loss at batch 8700: 0.001724, throughput: 24723 samples/sec Average loss at batch 8800: 0.001732, throughput: 24096 samples/sec Average loss at batch 8900: 0.001838, throughput: 24154 samples/sec Average loss at batch 9000: 0.002092, throughput: 27669 samples/sec Average loss at batch 9100: 0.002287, throughput: 25264 samples/sec Average loss at batch 9200: 0.002656, throughput: 25699 samples/sec Average loss at batch 9300: 0.002944, throughput: 26369 samples/sec Average loss at