t5_base-qg-ap-test
This model is a fine-tuned version of unicamp-dl/ptt5-base-portuguese-vocab on the tiagoblima/du-qg-squadv1_pt dataset. It achieves the following results on the evaluation set:
- Loss: 0.0163
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
- lr_scheduler_type: linear
- num_epochs: 100.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 1 | 12.8054 |
| No log | 2.0 | 2 | 10.7880 |
| No log | 3.0 | 3 | 8.8731 |
| No log | 4.0 | 4 | 7.4068 |
| No log | 5.0 | 5 | 6.4581 |
| No log | 6.0 | 6 | 5.6475 |
| No log | 7.0 | 7 | 4.9596 |
| No log | 8.0 | 8 | 4.5058 |
| No log | 9.0 | 9 | 4.0768 |
| No log | 10.0 | 10 | 3.7047 |
| No log | 11.0 | 11 | 3.4143 |
| No log | 12.0 | 12 | 3.1360 |
| No log | 13.0 | 13 | 2.8866 |
| No log | 14.0 | 14 | 2.6325 |
| No log | 15.0 | 15 | 2.3889 |
| No log | 16.0 | 16 | 2.1914 |
| No log | 17.0 | 17 | 2.0424 |
| No log | 18.0 | 18 | 1.9111 |
| No log | 19.0 | 19 | 1.7763 |
| No log | 20.0 | 20 | 1.6505 |
| No log | 21.0 | 21 | 1.5257 |
| No log | 22.0 | 22 | 1.4126 |
| No log | 23.0 | 23 | 1.3109 |
| No log | 24.0 | 24 | 1.2189 |
| No log | 25.0 | 25 | 1.1338 |
| No log | 26.0 | 26 | 1.0486 |
| No log | 27.0 | 27 | 0.9640 |
| No log | 28.0 | 28 | 0.8828 |
| No log | 29.0 | 29 | 0.8060 |
| No log | 30.0 | 30 | 0.7329 |
| No log | 31.0 | 31 | 0.6639 |
| No log | 32.0 | 32 | 0.6010 |
| No log | 33.0 | 33 | 0.5439 |
| No log | 34.0 | 34 | 0.4925 |
| No log | 35.0 | 35 | 0.4471 |
| No log | 36.0 | 36 | 0.4066 |
| No log | 37.0 | 37 | 0.3690 |
| No log | 38.0 | 38 | 0.3341 |
| No log | 39.0 | 39 | 0.3023 |
| No log | 40.0 | 40 | 0.2746 |
| No log | 41.0 | 41 | 0.2470 |
| No log | 42.0 | 42 | 0.2205 |
| No log | 43.0 | 43 | 0.1968 |
| No log | 44.0 | 44 | 0.1771 |
| No log | 45.0 | 45 | 0.1593 |
| No log | 46.0 | 46 | 0.1424 |
| No log | 47.0 | 47 | 0.1288 |
| No log | 48.0 | 48 | 0.1170 |
| No log | 49.0 | 49 | 0.1070 |
| No log | 50.0 | 50 | 0.0996 |
| No log | 51.0 | 51 | 0.0939 |
| No log | 52.0 | 52 | 0.0888 |
| No log | 53.0 | 53 | 0.0845 |
| No log | 54.0 | 54 | 0.0818 |
| No log | 55.0 | 55 | 0.0790 |
| No log | 56.0 | 56 | 0.0763 |
| No log | 57.0 | 57 | 0.0732 |
| No log | 58.0 | 58 | 0.0697 |
| No log | 59.0 | 59 | 0.0666 |
| No log | 60.0 | 60 | 0.0642 |
| No log | 61.0 | 61 | 0.0611 |
| No log | 62.0 | 62 | 0.0583 |
| No log | 63.0 | 63 | 0.0560 |
| No log | 64.0 | 64 | 0.0532 |
| No log | 65.0 | 65 | 0.0512 |
| No log | 66.0 | 66 | 0.0487 |
| No log | 67.0 | 67 | 0.0464 |
| No log | 68.0 | 68 | 0.0431 |
| No log | 69.0 | 69 | 0.0399 |
| No log | 70.0 | 70 | 0.0381 |
| No log | 71.0 | 71 | 0.0364 |
| No log | 72.0 | 72 | 0.0348 |
| No log | 73.0 | 73 | 0.0333 |
| No log | 74.0 | 74 | 0.0316 |
| No log | 75.0 | 75 | 0.0299 |
| No log | 76.0 | 76 | 0.0285 |
| No log | 77.0 | 77 | 0.0274 |
| No log | 78.0 | 78 | 0.0264 |
| No log | 79.0 | 79 | 0.0253 |
| No log | 80.0 | 80 | 0.0242 |
| No log | 81.0 | 81 | 0.0236 |
| No log | 82.0 | 82 | 0.0231 |
| No log | 83.0 | 83 | 0.0229 |
| No log | 84.0 | 84 | 0.0226 |
| No log | 85.0 | 85 | 0.0223 |
| No log | 86.0 | 86 | 0.0218 |
| No log | 87.0 | 87 | 0.0212 |
| No log | 88.0 | 88 | 0.0205 |
| No log | 89.0 | 89 | 0.0198 |
| No log | 90.0 | 90 | 0.0192 |
| No log | 91.0 | 91 | 0.0186 |
| No log | 92.0 | 92 | 0.0181 |
| No log | 93.0 | 93 | 0.0177 |
| No log | 94.0 | 94 | 0.0173 |
| No log | 95.0 | 95 | 0.0170 |
| No log | 96.0 | 96 | 0.0168 |
| No log | 97.0 | 97 | 0.0166 |
| No log | 98.0 | 98 | 0.0165 |
| No log | 99.0 | 99 | 0.0164 |
| 1.4009 | 100.0 | 100 | 0.0163 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.0
- Datasets 2.15.0
- Tokenizers 0.15.0