File size: 2,289 Bytes
c953d2d aee1fce bb628c8 0425676 feb9b8c 41cb6c1 95d0573 f82ecd8 4b9bc10 c373fb1 8f80131 0739091 a8a78b5 d1681df 3590960 124b390 fdfbf67 725bba3 788f6c0 ef9b400 c7bdc7a a2f3f71 b8bf0fc fccc8ea 69af903 4f842f1 adda4be 44a0120 92cb482 3f88d71 06ca727 fd68de2 2e91109 153c831 d6868ec 556658f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
Started at: 15:42:32
nb-bert-base, 5e-06, 128
({'_name_or_path': '/disk4/folder1/working/checkpoints/huggingface/native_pytorch/step4_8/', 'attention_probs_dropout_prob': 0.1, 'directionality': 'bidi', 'gradient_checkpointing': False, 'hidden_act': 'gelu', 'hidden_dropout_prob': 0.1, 'hidden_size': 768, 'initializer_range': 0.02, 'intermediate_size': 3072, 'layer_norm_eps': 1e-12, 'max_position_embeddings': 512, 'model_type': 'bert', 'num_attention_heads': 12, 'num_hidden_layers': 12, 'pad_token_id': 0, 'pooler_fc_size': 768, 'pooler_num_attention_heads': 12, 'pooler_num_fc_layers': 3, 'pooler_size_per_head': 128, 'pooler_type': 'first_token_transform', 'position_embedding_type': 'absolute', 'type_vocab_size': 2, 'vocab_size': 119547, '_commit_hash': '82b194c0b3ea1fcad65f1eceee04adb26f9f71ac'}, {})
Epoch: 0
Training loss: 0.33608432710170744 - MAE: 0.43243840742639095
Validation loss : 0.1794234970036675 - MAE: 0.32278704808536035
Epoch: 1
Training loss: 0.1674842369556427 - MAE: 0.31167954456795716
Validation loss : 0.1640064299983137 - MAE: 0.3063768331901366
Epoch: 2
Training loss: 0.1580326673388481 - MAE: 0.3004621400964959
Validation loss : 0.15081930116695516 - MAE: 0.2939987153720106
Epoch: 3
Training loss: 0.14433746680617332 - MAE: 0.2853955953382289
Validation loss : 0.13955745363936706 - MAE: 0.2805828900770382
Epoch: 4
Training loss: 0.12773224733769895 - MAE: 0.2682634582590704
Validation loss : 0.12777050628381617 - MAE: 0.26856869819032614
Epoch: 5
Training loss: 0.11507055133581162 - MAE: 0.2548698064814896
Validation loss : 0.11819195002317429 - MAE: 0.2559857516534626
Epoch: 6
Training loss: 0.10605520464479923 - MAE: 0.24685413996018704
Validation loss : 0.11385800119708567 - MAE: 0.2514239431805721
Epoch: 7
Training loss: 0.09987761311233044 - MAE: 0.23910935102620393
Validation loss : 0.10999320841887418 - MAE: 0.2490825267050983
Epoch: 8
Training loss: 0.09503072924911976 - MAE: 0.23424281710060202
Validation loss : 0.10815327263930265 - MAE: 0.24874807522133224
Epoch: 9
Training loss: 0.09119642566889524 - MAE: 0.229760940773865
Validation loss : 0.10649637790287242 - MAE: 0.2485144665264264
Epoch: 10
Training loss: 0.08799542726948857 - MAE: 0.2265422133769623
Validation loss : 0.10533954553744372 - MAE: 0.24686528907851468
|