Upload 6 files

final-model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6449d4d9b3888c8162027937aac17c089a102cb104e955dd269860fa572def9f
+size 463652197

loss.tsv

@@ -0,0 +1,11 @@
+EPOCH	TIMESTAMP	BAD_EPOCHS	LEARNING_RATE	TRAIN_LOSS	DEV_LOSS	DEV_PRECISION	DEV_RECALL	DEV_F1	DEV_ACCURACY
+1	19:57:09	0	0.0100	0.2008437729789472	0.09606283158063889	0.7451	0.7602	0.7526	0.6768
+2	20:09:48	0	0.0100	0.11846380610295361	0.07920133322477341	0.7996	0.8321	0.8155	0.7477
+3	20:21:56	0	0.0100	0.09812578978029159	0.07603894919157028	0.8276	0.8448	0.8361	0.7699
+4	20:37:53	1	0.0100	0.0853089421505088	0.07134225219488144	0.844	0.8235	0.8336	0.7731
+5	20:51:56	0	0.0100	0.07707492752456371	0.06873895972967148	0.8429	0.8531	0.848	0.7903
+6	21:04:16	0	0.0100	0.06804995682682495	0.05917559936642647	0.8827	0.8723	0.8775	0.8199
+7	21:19:27	0	0.0100	0.061280448328724924	0.061052411794662476	0.8729	0.8901	0.8814	0.8264
+8	21:31:53	1	0.0100	0.0552519113601074	0.06685522198677063	0.8813	0.8804	0.8808	0.8263
+9	21:44:13	0	0.0100	0.04966619410573325	0.057355064898729324	0.8888	0.8957	0.8922	0.8432
+10	22:02:22	1	0.0100	0.044047680323688256	0.06379110366106033	0.9023	0.8736	0.8877	0.8323

training.log

@@ -0,0 +1,525 @@
+2022-10-26 19:45:19,393 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,398 Model: "SequenceTagger(
+  (embeddings): TransformerWordEmbeddings(
+    (model): BertModel(
+      (embeddings): BertEmbeddings(
+        (word_embeddings): Embedding(35000, 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)
+      )
+      (encoder): BertEncoder(
+        (layer): ModuleList(
+          (0): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (1): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (2): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (3): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (4): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (5): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (6): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (7): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (8): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (9): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (10): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+              (dropout): Dropout(p=0.1, inplace=False)
+            )
+          )
+          (11): BertLayer(
+            (attention): BertAttention(
+              (self): BertSelfAttention(
+                (query): Linear(in_features=768, out_features=768, bias=True)
+                (key): Linear(in_features=768, out_features=768, bias=True)
+                (value): Linear(in_features=768, out_features=768, bias=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+              (output): BertSelfOutput(
+                (dense): Linear(in_features=768, out_features=768, bias=True)
+                (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
+                (dropout): Dropout(p=0.1, inplace=False)
+              )
+            )
+            (intermediate): BertIntermediate(
+              (dense): Linear(in_features=768, out_features=3072, bias=True)
+              (intermediate_act_fn): GELUActivation()
+            )
+            (output): BertOutput(
+              (dense): Linear(in_features=3072, out_features=768, bias=True)
+              (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)
+  (embedding2nn): Linear(in_features=768, out_features=768, bias=True)
+  (rnn): LSTM(768, 256, batch_first=True, bidirectional=True)
+  (linear): Linear(in_features=512, out_features=15, bias=True)
+  (loss_function): ViterbiLoss()
+  (crf): CRF()
+)"
+2022-10-26 19:45:19,409 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,415 Corpus: "Corpus: 8551 train + 1425 dev + 1425 test sentences"
+2022-10-26 19:45:19,418 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,425 Parameters:
+2022-10-26 19:45:19,429  - learning_rate: "0.010000"
+2022-10-26 19:45:19,436  - mini_batch_size: "8"
+2022-10-26 19:45:19,441  - patience: "3"
+2022-10-26 19:45:19,446  - anneal_factor: "0.5"
+2022-10-26 19:45:19,447  - max_epochs: "10"
+2022-10-26 19:45:19,466  - shuffle: "True"
+2022-10-26 19:45:19,470  - train_with_dev: "False"
+2022-10-26 19:45:19,475  - batch_growth_annealing: "False"
+2022-10-26 19:45:19,476 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,479 Model training base path: "/content/model/mono_ner"
+2022-10-26 19:45:19,480 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,484 Device: cuda:0
+2022-10-26 19:45:19,489 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:45:19,491 Embeddings storage mode: none
+2022-10-26 19:45:19,496 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:46:27,364 epoch 1 - iter 106/1069 - loss 0.49979466 - samples/sec: 12.50 - lr: 0.010000
+2022-10-26 19:47:29,408 epoch 1 - iter 212/1069 - loss 0.36858293 - samples/sec: 13.67 - lr: 0.010000
+2022-10-26 19:48:32,710 epoch 1 - iter 318/1069 - loss 0.31288040 - samples/sec: 13.40 - lr: 0.010000
+2022-10-26 19:49:36,271 epoch 1 - iter 424/1069 - loss 0.27906252 - samples/sec: 13.34 - lr: 0.010000
+2022-10-26 19:50:40,278 epoch 1 - iter 530/1069 - loss 0.25802546 - samples/sec: 13.25 - lr: 0.010000
+2022-10-26 19:51:45,008 epoch 1 - iter 636/1069 - loss 0.24111842 - samples/sec: 13.10 - lr: 0.010000
+2022-10-26 19:52:47,602 epoch 1 - iter 742/1069 - loss 0.22829427 - samples/sec: 13.55 - lr: 0.010000
+2022-10-26 19:53:50,115 epoch 1 - iter 848/1069 - loss 0.21731094 - samples/sec: 13.57 - lr: 0.010000
+2022-10-26 19:54:53,793 epoch 1 - iter 954/1069 - loss 0.20876564 - samples/sec: 13.32 - lr: 0.010000
+2022-10-26 19:55:55,252 epoch 1 - iter 1060/1069 - loss 0.20166716 - samples/sec: 13.80 - lr: 0.010000
+2022-10-26 19:56:00,400 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:56:00,402 EPOCH 1 done: loss 0.2008 - lr 0.010000
+2022-10-26 19:57:09,701 Evaluating as a multi-label problem: False
+2022-10-26 19:57:09,740 DEV : loss 0.09606283158063889 - f1-score (micro avg)  0.7526
+2022-10-26 19:57:09,783 BAD EPOCHS (no improvement): 0
+2022-10-26 19:57:09,785 saving best model
+2022-10-26 19:57:11,433 ----------------------------------------------------------------------------------------------------
+2022-10-26 19:58:18,467 epoch 2 - iter 106/1069 - loss 0.12276787 - samples/sec: 12.65 - lr: 0.010000
+2022-10-26 19:59:24,322 epoch 2 - iter 212/1069 - loss 0.12231755 - samples/sec: 12.88 - lr: 0.010000
+2022-10-26 20:00:41,700 epoch 2 - iter 318/1069 - loss 0.12435630 - samples/sec: 10.96 - lr: 0.010000
+2022-10-26 20:01:46,059 epoch 2 - iter 424/1069 - loss 0.12564768 - samples/sec: 13.18 - lr: 0.010000
+2022-10-26 20:02:49,678 epoch 2 - iter 530/1069 - loss 0.12512958 - samples/sec: 13.33 - lr: 0.010000
+2022-10-26 20:04:05,654 epoch 2 - iter 636/1069 - loss 0.12238487 - samples/sec: 11.16 - lr: 0.010000
+2022-10-26 20:05:09,552 epoch 2 - iter 742/1069 - loss 0.12010170 - samples/sec: 13.27 - lr: 0.010000
+2022-10-26 20:06:14,022 epoch 2 - iter 848/1069 - loss 0.11967127 - samples/sec: 13.16 - lr: 0.010000
+2022-10-26 20:07:19,659 epoch 2 - iter 954/1069 - loss 0.11888882 - samples/sec: 12.92 - lr: 0.010000
+2022-10-26 20:08:29,253 epoch 2 - iter 1060/1069 - loss 0.11866747 - samples/sec: 12.19 - lr: 0.010000
+2022-10-26 20:08:34,370 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:08:34,372 EPOCH 2 done: loss 0.1185 - lr 0.010000
+2022-10-26 20:09:47,920 Evaluating as a multi-label problem: False
+2022-10-26 20:09:47,955 DEV : loss 0.07920133322477341 - f1-score (micro avg)  0.8155
+2022-10-26 20:09:47,998 BAD EPOCHS (no improvement): 0
+2022-10-26 20:09:48,000 saving best model
+2022-10-26 20:09:49,587 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:10:53,964 epoch 3 - iter 106/1069 - loss 0.10166018 - samples/sec: 13.18 - lr: 0.010000
+2022-10-26 20:11:56,797 epoch 3 - iter 212/1069 - loss 0.10111216 - samples/sec: 13.50 - lr: 0.010000
+2022-10-26 20:13:03,180 epoch 3 - iter 318/1069 - loss 0.10239146 - samples/sec: 12.78 - lr: 0.010000
+2022-10-26 20:14:08,543 epoch 3 - iter 424/1069 - loss 0.10173990 - samples/sec: 12.98 - lr: 0.010000
+2022-10-26 20:15:13,145 epoch 3 - iter 530/1069 - loss 0.10135509 - samples/sec: 13.13 - lr: 0.010000
+2022-10-26 20:16:19,356 epoch 3 - iter 636/1069 - loss 0.10020505 - samples/sec: 12.81 - lr: 0.010000
+2022-10-26 20:17:21,470 epoch 3 - iter 742/1069 - loss 0.10033292 - samples/sec: 13.65 - lr: 0.010000
+2022-10-26 20:18:25,712 epoch 3 - iter 848/1069 - loss 0.09965180 - samples/sec: 13.20 - lr: 0.010000
+2022-10-26 20:19:32,123 epoch 3 - iter 954/1069 - loss 0.09942363 - samples/sec: 12.77 - lr: 0.010000
+2022-10-26 20:20:37,362 epoch 3 - iter 1060/1069 - loss 0.09818458 - samples/sec: 13.00 - lr: 0.010000
+2022-10-26 20:20:42,922 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:20:42,923 EPOCH 3 done: loss 0.0981 - lr 0.010000
+2022-10-26 20:21:56,678 Evaluating as a multi-label problem: False
+2022-10-26 20:21:56,717 DEV : loss 0.07603894919157028 - f1-score (micro avg)  0.8361
+2022-10-26 20:21:56,759 BAD EPOCHS (no improvement): 0
+2022-10-26 20:21:56,761 saving best model
+2022-10-26 20:21:58,329 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:23:02,865 epoch 4 - iter 106/1069 - loss 0.08581557 - samples/sec: 13.14 - lr: 0.010000
+2022-10-26 20:24:06,558 epoch 4 - iter 212/1069 - loss 0.08690126 - samples/sec: 13.32 - lr: 0.010000
+2022-10-26 20:25:11,549 epoch 4 - iter 318/1069 - loss 0.08740134 - samples/sec: 13.05 - lr: 0.010000
+2022-10-26 20:26:16,171 epoch 4 - iter 424/1069 - loss 0.08691255 - samples/sec: 13.12 - lr: 0.010000
+2022-10-26 20:27:21,108 epoch 4 - iter 530/1069 - loss 0.08743159 - samples/sec: 13.06 - lr: 0.010000
+2022-10-26 20:28:26,306 epoch 4 - iter 636/1069 - loss 0.08700733 - samples/sec: 13.01 - lr: 0.010000
+2022-10-26 20:29:28,907 epoch 4 - iter 742/1069 - loss 0.08700591 - samples/sec: 13.55 - lr: 0.010000
+2022-10-26 20:30:34,735 epoch 4 - iter 848/1069 - loss 0.08615337 - samples/sec: 12.88 - lr: 0.010000
+2022-10-26 20:32:03,266 epoch 4 - iter 954/1069 - loss 0.08562659 - samples/sec: 9.58 - lr: 0.010000
+2022-10-26 20:33:59,270 epoch 4 - iter 1060/1069 - loss 0.08544457 - samples/sec: 7.31 - lr: 0.010000
+2022-10-26 20:34:09,369 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:34:09,371 EPOCH 4 done: loss 0.0853 - lr 0.010000
+2022-10-26 20:37:53,248 Evaluating as a multi-label problem: False
+2022-10-26 20:37:53,283 DEV : loss 0.07134225219488144 - f1-score (micro avg)  0.8336
+2022-10-26 20:37:53,326 BAD EPOCHS (no improvement): 1
+2022-10-26 20:37:53,328 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:39:45,902 epoch 5 - iter 106/1069 - loss 0.07612726 - samples/sec: 7.53 - lr: 0.010000
+2022-10-26 20:41:42,470 epoch 5 - iter 212/1069 - loss 0.07932025 - samples/sec: 7.28 - lr: 0.010000
+2022-10-26 20:43:01,451 epoch 5 - iter 318/1069 - loss 0.07766485 - samples/sec: 10.74 - lr: 0.010000
+2022-10-26 20:44:06,242 epoch 5 - iter 424/1069 - loss 0.07782655 - samples/sec: 13.09 - lr: 0.010000
+2022-10-26 20:45:10,011 epoch 5 - iter 530/1069 - loss 0.07797363 - samples/sec: 13.30 - lr: 0.010000
+2022-10-26 20:46:18,444 epoch 5 - iter 636/1069 - loss 0.07784710 - samples/sec: 12.39 - lr: 0.010000
+2022-10-26 20:47:22,712 epoch 5 - iter 742/1069 - loss 0.07764170 - samples/sec: 13.20 - lr: 0.010000
+2022-10-26 20:48:26,544 epoch 5 - iter 848/1069 - loss 0.07765970 - samples/sec: 13.29 - lr: 0.010000
+2022-10-26 20:49:32,065 epoch 5 - iter 954/1069 - loss 0.07726613 - samples/sec: 12.94 - lr: 0.010000
+2022-10-26 20:50:36,714 epoch 5 - iter 1060/1069 - loss 0.07692019 - samples/sec: 13.12 - lr: 0.010000
+2022-10-26 20:50:41,823 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:50:41,825 EPOCH 5 done: loss 0.0771 - lr 0.010000
+2022-10-26 20:51:56,635 Evaluating as a multi-label problem: False
+2022-10-26 20:51:56,681 DEV : loss 0.06873895972967148 - f1-score (micro avg)  0.848
+2022-10-26 20:51:56,730 BAD EPOCHS (no improvement): 0
+2022-10-26 20:51:56,732 saving best model
+2022-10-26 20:51:58,276 ----------------------------------------------------------------------------------------------------
+2022-10-26 20:53:04,269 epoch 6 - iter 106/1069 - loss 0.07259857 - samples/sec: 12.85 - lr: 0.010000
+2022-10-26 20:54:08,435 epoch 6 - iter 212/1069 - loss 0.06894409 - samples/sec: 13.22 - lr: 0.010000
+2022-10-26 20:55:15,290 epoch 6 - iter 318/1069 - loss 0.06918623 - samples/sec: 12.69 - lr: 0.010000
+2022-10-26 20:56:20,441 epoch 6 - iter 424/1069 - loss 0.06917844 - samples/sec: 13.02 - lr: 0.010000
+2022-10-26 20:57:24,834 epoch 6 - iter 530/1069 - loss 0.06940973 - samples/sec: 13.17 - lr: 0.010000
+2022-10-26 20:58:31,661 epoch 6 - iter 636/1069 - loss 0.06932249 - samples/sec: 12.69 - lr: 0.010000
+2022-10-26 20:59:37,057 epoch 6 - iter 742/1069 - loss 0.06858729 - samples/sec: 12.97 - lr: 0.010000
+2022-10-26 21:00:42,037 epoch 6 - iter 848/1069 - loss 0.06850174 - samples/sec: 13.05 - lr: 0.010000
+2022-10-26 21:01:48,234 epoch 6 - iter 954/1069 - loss 0.06855966 - samples/sec: 12.81 - lr: 0.010000
+2022-10-26 21:02:54,530 epoch 6 - iter 1060/1069 - loss 0.06812598 - samples/sec: 12.79 - lr: 0.010000
+2022-10-26 21:03:00,480 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:03:00,482 EPOCH 6 done: loss 0.0680 - lr 0.010000
+2022-10-26 21:04:16,435 Evaluating as a multi-label problem: False
+2022-10-26 21:04:16,476 DEV : loss 0.05917559936642647 - f1-score (micro avg)  0.8775
+2022-10-26 21:04:16,522 BAD EPOCHS (no improvement): 0
+2022-10-26 21:04:16,526 saving best model
+2022-10-26 21:04:18,071 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:05:24,303 epoch 7 - iter 106/1069 - loss 0.06352705 - samples/sec: 12.81 - lr: 0.010000
+2022-10-26 21:06:30,784 epoch 7 - iter 212/1069 - loss 0.06166309 - samples/sec: 12.76 - lr: 0.010000
+2022-10-26 21:07:35,118 epoch 7 - iter 318/1069 - loss 0.06134693 - samples/sec: 13.18 - lr: 0.010000
+2022-10-26 21:08:39,228 epoch 7 - iter 424/1069 - loss 0.06161759 - samples/sec: 13.23 - lr: 0.010000
+2022-10-26 21:10:15,880 epoch 7 - iter 530/1069 - loss 0.06137938 - samples/sec: 8.77 - lr: 0.010000
+2022-10-26 21:12:14,808 epoch 7 - iter 636/1069 - loss 0.06149529 - samples/sec: 7.13 - lr: 0.010000
+2022-10-26 21:14:13,856 epoch 7 - iter 742/1069 - loss 0.06173201 - samples/sec: 7.12 - lr: 0.010000
+2022-10-26 21:15:51,294 epoch 7 - iter 848/1069 - loss 0.06166752 - samples/sec: 8.70 - lr: 0.010000
+2022-10-26 21:16:59,785 epoch 7 - iter 954/1069 - loss 0.06152770 - samples/sec: 12.38 - lr: 0.010000
+2022-10-26 21:18:05,005 epoch 7 - iter 1060/1069 - loss 0.06131402 - samples/sec: 13.00 - lr: 0.010000
+2022-10-26 21:18:10,767 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:18:10,769 EPOCH 7 done: loss 0.0613 - lr 0.010000
+2022-10-26 21:19:27,868 Evaluating as a multi-label problem: False
+2022-10-26 21:19:27,905 DEV : loss 0.061052411794662476 - f1-score (micro avg)  0.8814
+2022-10-26 21:19:27,952 BAD EPOCHS (no improvement): 0
+2022-10-26 21:19:27,954 saving best model
+2022-10-26 21:19:29,378 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:20:36,789 epoch 8 - iter 106/1069 - loss 0.05390116 - samples/sec: 12.58 - lr: 0.010000
+2022-10-26 21:21:41,786 epoch 8 - iter 212/1069 - loss 0.05771654 - samples/sec: 13.05 - lr: 0.010000
+2022-10-26 21:22:48,800 epoch 8 - iter 318/1069 - loss 0.05630827 - samples/sec: 12.66 - lr: 0.010000
+2022-10-26 21:23:54,308 epoch 8 - iter 424/1069 - loss 0.05571937 - samples/sec: 12.95 - lr: 0.010000
+2022-10-26 21:25:00,994 epoch 8 - iter 530/1069 - loss 0.05600622 - samples/sec: 12.72 - lr: 0.010000
+2022-10-26 21:26:05,543 epoch 8 - iter 636/1069 - loss 0.05638838 - samples/sec: 13.14 - lr: 0.010000
+2022-10-26 21:27:11,826 epoch 8 - iter 742/1069 - loss 0.05616568 - samples/sec: 12.80 - lr: 0.010000
+2022-10-26 21:28:18,954 epoch 8 - iter 848/1069 - loss 0.05584409 - samples/sec: 12.64 - lr: 0.010000
+2022-10-26 21:29:25,542 epoch 8 - iter 954/1069 - loss 0.05561947 - samples/sec: 12.74 - lr: 0.010000
+2022-10-26 21:30:30,533 epoch 8 - iter 1060/1069 - loss 0.05524983 - samples/sec: 13.05 - lr: 0.010000
+2022-10-26 21:30:35,751 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:30:35,755 EPOCH 8 done: loss 0.0553 - lr 0.010000
+2022-10-26 21:31:53,000 Evaluating as a multi-label problem: False
+2022-10-26 21:31:53,038 DEV : loss 0.06685522198677063 - f1-score (micro avg)  0.8808
+2022-10-26 21:31:53,088 BAD EPOCHS (no improvement): 1
+2022-10-26 21:31:53,092 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:33:00,202 epoch 9 - iter 106/1069 - loss 0.04591263 - samples/sec: 12.64 - lr: 0.010000
+2022-10-26 21:34:05,608 epoch 9 - iter 212/1069 - loss 0.04753505 - samples/sec: 12.97 - lr: 0.010000
+2022-10-26 21:35:08,841 epoch 9 - iter 318/1069 - loss 0.04983626 - samples/sec: 13.41 - lr: 0.010000
+2022-10-26 21:36:15,599 epoch 9 - iter 424/1069 - loss 0.04851610 - samples/sec: 12.70 - lr: 0.010000
+2022-10-26 21:37:22,043 epoch 9 - iter 530/1069 - loss 0.04882362 - samples/sec: 12.77 - lr: 0.010000
+2022-10-26 21:38:26,514 epoch 9 - iter 636/1069 - loss 0.04925004 - samples/sec: 13.16 - lr: 0.010000
+2022-10-26 21:39:34,184 epoch 9 - iter 742/1069 - loss 0.04945580 - samples/sec: 12.53 - lr: 0.010000
+2022-10-26 21:40:39,778 epoch 9 - iter 848/1069 - loss 0.04945835 - samples/sec: 12.93 - lr: 0.010000
+2022-10-26 21:41:44,710 epoch 9 - iter 954/1069 - loss 0.04953811 - samples/sec: 13.06 - lr: 0.010000
+2022-10-26 21:42:52,682 epoch 9 - iter 1060/1069 - loss 0.04944091 - samples/sec: 12.48 - lr: 0.010000
+2022-10-26 21:42:57,825 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:42:57,826 EPOCH 9 done: loss 0.0497 - lr 0.010000
+2022-10-26 21:44:13,770 Evaluating as a multi-label problem: False
+2022-10-26 21:44:13,809 DEV : loss 0.057355064898729324 - f1-score (micro avg)  0.8922
+2022-10-26 21:44:13,856 BAD EPOCHS (no improvement): 0
+2022-10-26 21:44:13,859 saving best model
+2022-10-26 21:44:15,333 ----------------------------------------------------------------------------------------------------
+2022-10-26 21:45:22,992 epoch 10 - iter 106/1069 - loss 0.03999971 - samples/sec: 12.54 - lr: 0.010000
+2022-10-26 21:46:28,166 epoch 10 - iter 212/1069 - loss 0.04223290 - samples/sec: 13.01 - lr: 0.010000
+2022-10-26 21:47:34,530 epoch 10 - iter 318/1069 - loss 0.04233629 - samples/sec: 12.78 - lr: 0.010000
+2022-10-26 21:49:21,523 epoch 10 - iter 424/1069 - loss 0.04293457 - samples/sec: 7.93 - lr: 0.010000
+2022-10-26 21:51:20,933 epoch 10 - iter 530/1069 - loss 0.04261612 - samples/sec: 7.10 - lr: 0.010000
+2022-10-26 21:53:16,486 epoch 10 - iter 636/1069 - loss 0.04316492 - samples/sec: 7.34 - lr: 0.010000
+2022-10-26 21:55:14,355 epoch 10 - iter 742/1069 - loss 0.04313719 - samples/sec: 7.20 - lr: 0.010000
+2022-10-26 21:57:14,471 epoch 10 - iter 848/1069 - loss 0.04345674 - samples/sec: 7.06 - lr: 0.010000
+2022-10-26 21:59:14,125 epoch 10 - iter 954/1069 - loss 0.04368164 - samples/sec: 7.09 - lr: 0.010000
+2022-10-26 22:01:02,494 epoch 10 - iter 1060/1069 - loss 0.04413420 - samples/sec: 7.83 - lr: 0.010000
+2022-10-26 22:01:08,438 ----------------------------------------------------------------------------------------------------
+2022-10-26 22:01:08,440 EPOCH 10 done: loss 0.0440 - lr 0.010000
+2022-10-26 22:02:22,434 Evaluating as a multi-label problem: False
+2022-10-26 22:02:22,472 DEV : loss 0.06379110366106033 - f1-score (micro avg)  0.8877
+2022-10-26 22:02:22,522 BAD EPOCHS (no improvement): 1
+2022-10-26 22:02:23,953 ----------------------------------------------------------------------------------------------------
+2022-10-26 22:02:23,963 loading file /content/model/mono_ner/best-model.pt
+2022-10-26 22:02:26,538 SequenceTagger predicts: Dictionary with 15 tags: O, S-PER, B-PER, E-PER, I-PER, S-MISC, B-MISC, E-MISC, I-MISC, S-LOC, B-LOC, E-LOC, I-LOC, <START>, <STOP>
+2022-10-26 22:03:39,014 Evaluating as a multi-label problem: False
+2022-10-26 22:03:39,054 0.8798	0.8959	0.8878	0.8324
+2022-10-26 22:03:39,056 
+Results:
+- F-score (micro) 0.8878
+- F-score (macro) 0.8574
+- Accuracy 0.8324
+
+By class:
+              precision    recall  f1-score   support
+
+         PER     0.9124    0.9445    0.9282      2127
+        MISC     0.8092    0.8317    0.8203       933
+         LOC     0.8686    0.7835    0.8238       388
+
+   micro avg     0.8798    0.8959    0.8878      3448
+   macro avg     0.8634    0.8533    0.8574      3448
+weighted avg     0.8795    0.8959    0.8872      3448
+
+2022-10-26 22:03:39,059 ----------------------------------------------------------------------------------------------------