scenario-NON-KD-PO-COPY-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_
This model is a fine-tuned version of haryoaw/scenario-MDBT-TCR-TSM on the tweet_sentiment_multilingual dataset. It achieves the following results on the evaluation set:
- Loss: 5.1975
- Accuracy: 0.5644
- F1: 0.5639
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 55
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
---|---|---|---|---|---|
0.9766 | 1.0870 | 500 | 0.9668 | 0.5679 | 0.5672 |
0.7795 | 2.1739 | 1000 | 0.9971 | 0.5880 | 0.5855 |
0.6025 | 3.2609 | 1500 | 1.2126 | 0.5760 | 0.5682 |
0.4143 | 4.3478 | 2000 | 1.3620 | 0.5733 | 0.5720 |
0.288 | 5.4348 | 2500 | 1.7279 | 0.5644 | 0.5624 |
0.2004 | 6.5217 | 3000 | 2.0524 | 0.5617 | 0.5629 |
0.1439 | 7.6087 | 3500 | 2.4904 | 0.5590 | 0.5594 |
0.1197 | 8.6957 | 4000 | 2.3080 | 0.5606 | 0.5581 |
0.1096 | 9.7826 | 4500 | 2.6392 | 0.5667 | 0.5629 |
0.0904 | 10.8696 | 5000 | 2.8438 | 0.5478 | 0.5498 |
0.0783 | 11.9565 | 5500 | 2.9731 | 0.5625 | 0.5558 |
0.0617 | 13.0435 | 6000 | 3.5176 | 0.5586 | 0.5596 |
0.0571 | 14.1304 | 6500 | 3.5156 | 0.5644 | 0.5657 |
0.0524 | 15.2174 | 7000 | 3.1091 | 0.5594 | 0.5574 |
0.0535 | 16.3043 | 7500 | 2.9773 | 0.5664 | 0.5634 |
0.0423 | 17.3913 | 8000 | 3.6352 | 0.5633 | 0.5641 |
0.0385 | 18.4783 | 8500 | 3.7201 | 0.5675 | 0.5647 |
0.0372 | 19.5652 | 9000 | 4.0422 | 0.5625 | 0.5599 |
0.0332 | 20.6522 | 9500 | 3.5064 | 0.5706 | 0.5708 |
0.0293 | 21.7391 | 10000 | 4.0279 | 0.5617 | 0.5617 |
0.0192 | 22.8261 | 10500 | 4.4137 | 0.5671 | 0.5637 |
0.0278 | 23.9130 | 11000 | 4.0800 | 0.5621 | 0.5599 |
0.0239 | 25.0 | 11500 | 3.9079 | 0.5606 | 0.5599 |
0.0221 | 26.0870 | 12000 | 4.1928 | 0.5702 | 0.5672 |
0.0164 | 27.1739 | 12500 | 4.3024 | 0.5586 | 0.5524 |
0.0155 | 28.2609 | 13000 | 4.4464 | 0.5660 | 0.5659 |
0.0206 | 29.3478 | 13500 | 4.4741 | 0.5579 | 0.5569 |
0.0153 | 30.4348 | 14000 | 4.2231 | 0.5505 | 0.5514 |
0.0112 | 31.5217 | 14500 | 4.4476 | 0.5644 | 0.5620 |
0.0119 | 32.6087 | 15000 | 4.4276 | 0.5583 | 0.5561 |
0.0109 | 33.6957 | 15500 | 4.4156 | 0.5621 | 0.5617 |
0.0113 | 34.7826 | 16000 | 4.0354 | 0.5633 | 0.5621 |
0.0084 | 35.8696 | 16500 | 4.7380 | 0.5590 | 0.5568 |
0.0085 | 36.9565 | 17000 | 4.3942 | 0.5644 | 0.5632 |
0.01 | 38.0435 | 17500 | 4.4828 | 0.5687 | 0.5682 |
0.0056 | 39.1304 | 18000 | 4.7518 | 0.5640 | 0.5605 |
0.0041 | 40.2174 | 18500 | 4.8487 | 0.5725 | 0.5719 |
0.0058 | 41.3043 | 19000 | 4.5515 | 0.5698 | 0.5701 |
0.0044 | 42.3913 | 19500 | 4.9174 | 0.5640 | 0.5630 |
0.0049 | 43.4783 | 20000 | 4.8322 | 0.5664 | 0.5659 |
0.004 | 44.5652 | 20500 | 4.8014 | 0.5660 | 0.5656 |
0.0008 | 45.6522 | 21000 | 5.1207 | 0.5644 | 0.5647 |
0.0033 | 46.7391 | 21500 | 5.1209 | 0.5637 | 0.5636 |
0.002 | 47.8261 | 22000 | 5.1817 | 0.5610 | 0.5605 |
0.0012 | 48.9130 | 22500 | 5.2011 | 0.5640 | 0.5630 |
0.0022 | 50.0 | 23000 | 5.1975 | 0.5644 | 0.5639 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.1.1+cu121
- Datasets 2.14.5
- Tokenizers 0.19.1
- Downloads last month
- 5
Model tree for haryoaw/scenario-NON-KD-PO-COPY-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_
Base model
microsoft/mdeberta-v3-base
Finetuned
haryoaw/scenario-MDBT-TCR-TSM