Flair-Persian-NER / training.log
PooryaPiroozfar's picture
new version of Persian NER
5607485
2023-01-08 08:23:21,495 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,498 Model: "SequenceTagger(
(embeddings): TransformerWordEmbeddings(
(model): BertModel(
(embeddings): BertEmbeddings(
(word_embeddings): Embedding(100000, 768, padding_idx=0)
(position_embeddings): Embedding(512, 768)
(token_type_embeddings): Embedding(2, 768)
(LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(pooler): BertPooler(
(dense): Linear(in_features=768, out_features=768, bias=True)
(activation): Tanh()
)
)
(word_dropout): WordDropout(p=0.05)
(locked_dropout): LockedDropout(p=0.5)
(linear): Linear(in_features=768, out_features=18, bias=True)
(beta): 1.0
(weights): None
(weight_tensor) None
)"
2023-01-08 08:23:21,500 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,505 Corpus: "Corpus: 26116 train + 2902 dev + 1572 test sentences"
2023-01-08 08:23:21,506 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,506 Parameters:
2023-01-08 08:23:21,507 - learning_rate: "5e-06"
2023-01-08 08:23:21,509 - mini_batch_size: "4"
2023-01-08 08:23:21,510 - patience: "3"
2023-01-08 08:23:21,512 - anneal_factor: "0.5"
2023-01-08 08:23:21,513 - max_epochs: "25"
2023-01-08 08:23:21,513 - shuffle: "False"
2023-01-08 08:23:21,514 - train_with_dev: "False"
2023-01-08 08:23:21,515 - batch_growth_annealing: "False"
2023-01-08 08:23:21,516 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,517 Model training base path: "resources/taggers/NSURL-2019_25epochs"
2023-01-08 08:23:21,518 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,519 Device: cuda:0
2023-01-08 08:23:21,519 ----------------------------------------------------------------------------------------------------
2023-01-08 08:23:21,520 Embeddings storage mode: none
2023-01-08 18:00:13,690 ----------------------------------------------------------------------------------------------------
2023-01-08 18:02:30,863 epoch 25 - iter 652/6529 - loss 0.12185023 - samples/sec: 19.02 - lr: 0.000000
2023-01-08 18:04:48,105 epoch 25 - iter 1304/6529 - loss 0.12151675 - samples/sec: 19.01 - lr: 0.000000
2023-01-08 18:07:03,845 epoch 25 - iter 1956/6529 - loss 0.12293666 - samples/sec: 19.22 - lr: 0.000000
2023-01-08 18:09:20,797 epoch 25 - iter 2608/6529 - loss 0.12248209 - samples/sec: 19.05 - lr: 0.000000
2023-01-08 18:11:38,782 epoch 25 - iter 3260/6529 - loss 0.12236612 - samples/sec: 18.91 - lr: 0.000000
2023-01-08 18:13:57,739 epoch 25 - iter 3912/6529 - loss 0.12284535 - samples/sec: 18.78 - lr: 0.000000
2023-01-08 18:16:19,460 epoch 25 - iter 4564/6529 - loss 0.12312537 - samples/sec: 18.41 - lr: 0.000000
2023-01-08 18:18:34,844 epoch 25 - iter 5216/6529 - loss 0.12315613 - samples/sec: 19.27 - lr: 0.000000
2023-01-08 18:20:52,724 epoch 25 - iter 5868/6529 - loss 0.12280164 - samples/sec: 18.92 - lr: 0.000000
2023-01-08 18:23:11,733 epoch 25 - iter 6520/6529 - loss 0.12286952 - samples/sec: 18.77 - lr: 0.000000
2023-01-08 18:23:13,587 ----------------------------------------------------------------------------------------------------
2023-01-08 18:23:13,590 EPOCH 25 done: loss 0.1229 - lr 0.0000000
2023-01-08 18:24:28,587 DEV : loss 0.1607247292995453 - f1-score (micro avg) 0.9119
2023-01-08 18:24:28,641 BAD EPOCHS (no improvement): 4
2023-01-08 18:24:29,854 ----------------------------------------------------------------------------------------------------
2023-01-08 18:24:29,857 Testing using last state of model ...
2023-01-08 18:25:11,654 0.9081 0.8984 0.9033 0.8277
2023-01-08 18:25:11,656
Results:
- F-score (micro) 0.9033
- F-score (macro) 0.8976
- Accuracy 0.8277
By class:
precision recall f1-score support
ORG 0.9016 0.8667 0.8838 1523
LOC 0.9113 0.9305 0.9208 1425
PER 0.9216 0.9322 0.9269 1224
DAT 0.8623 0.7958 0.8277 480
MON 0.9665 0.9558 0.9611 181
PCT 0.9375 0.9740 0.9554 77
TIM 0.8235 0.7925 0.8077 53
micro avg 0.9081 0.8984 0.9033 4963
macro avg 0.9035 0.8925 0.8976 4963
weighted avg 0.9076 0.8984 0.9028 4963
samples avg 0.8277 0.8277 0.8277 4963
2023-01-08 18:25:11,656 ----------------------------------------------------------------------------------------------------