<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
mt5-base-fce-e8-b16
This model is a fine-tuned version of google/mt5-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3758
- Rouge1: 84.5938
- Rouge2: 76.5987
- Rougel: 84.0063
- Rougelsum: 84.0286
- Gen Len: 15.4865
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: 0.001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adafactor
- lr_scheduler_type: linear
- num_epochs: 8
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.5646 | 0.23 | 400 | 0.5403 | 83.2786 | 74.549 | 82.6906 | 82.6978 | 15.5126 |
| 0.6122 | 0.45 | 800 | 0.4896 | 84.3453 | 75.5159 | 83.7564 | 83.7691 | 15.4500 |
| 0.5041 | 0.68 | 1200 | 0.4294 | 84.2563 | 75.8731 | 83.6118 | 83.6071 | 15.4760 |
| 0.4594 | 0.9 | 1600 | 0.4136 | 84.7369 | 76.6048 | 84.1541 | 84.1573 | 15.4651 |
| 0.3861 | 1.13 | 2000 | 0.4121 | 84.6947 | 76.574 | 84.0885 | 84.095 | 15.4642 |
| 0.3382 | 1.35 | 2400 | 0.3899 | 84.5537 | 76.4381 | 83.9421 | 83.951 | 15.4651 |
| 0.3442 | 1.58 | 2800 | 0.3866 | 84.6272 | 76.6256 | 84.0616 | 84.0804 | 15.4674 |
| 0.3388 | 1.81 | 3200 | 0.3758 | 84.5938 | 76.5987 | 84.0063 | 84.0286 | 15.4865 |
| 0.3109 | 2.03 | 3600 | 0.3822 | 84.5223 | 76.5703 | 83.9217 | 83.9438 | 15.4710 |
| 0.2254 | 2.26 | 4000 | 0.3923 | 84.3225 | 76.4146 | 83.7686 | 83.7789 | 15.4596 |
| 0.236 | 2.48 | 4400 | 0.3932 | 84.4412 | 76.4434 | 83.8515 | 83.8815 | 15.4692 |
| 0.2395 | 2.71 | 4800 | 0.3849 | 84.2211 | 76.3678 | 83.6444 | 83.6462 | 15.4614 |
| 0.2458 | 2.93 | 5200 | 0.3850 | 84.3534 | 76.598 | 83.8321 | 83.8366 | 15.4587 |
| 0.1832 | 3.16 | 5600 | 0.3973 | 84.4197 | 76.7844 | 83.8758 | 83.8781 | 15.4678 |
| 0.1576 | 3.39 | 6000 | 0.4082 | 84.1841 | 76.4425 | 83.6272 | 83.618 | 15.4783 |
| 0.1635 | 3.61 | 6400 | 0.3996 | 84.2051 | 76.3261 | 83.6613 | 83.6599 | 15.4788 |
| 0.1667 | 3.84 | 6800 | 0.3940 | 84.4538 | 76.8139 | 83.8887 | 83.8886 | 15.4610 |
| 0.145 | 4.06 | 7200 | 0.4260 | 84.4028 | 76.8101 | 83.8844 | 83.8824 | 15.4628 |
| 0.107 | 4.29 | 7600 | 0.4403 | 84.3559 | 76.8066 | 83.8048 | 83.807 | 15.4587 |
| 0.1078 | 4.51 | 8000 | 0.4337 | 84.3045 | 76.8011 | 83.7587 | 83.7699 | 15.4742 |
| 0.1114 | 4.74 | 8400 | 0.4334 | 84.2865 | 76.5415 | 83.7221 | 83.718 | 15.4820 |
| 0.1104 | 4.97 | 8800 | 0.4273 | 84.3211 | 76.8211 | 83.7795 | 83.7726 | 15.4838 |
| 0.0732 | 5.19 | 9200 | 0.4787 | 84.3459 | 76.752 | 83.777 | 83.7552 | 15.4829 |
| 0.069 | 5.42 | 9600 | 0.4839 | 84.4351 | 76.8848 | 83.8682 | 83.8584 | 15.4811 |
| 0.0713 | 5.64 | 10000 | 0.4896 | 84.2962 | 76.7428 | 83.7387 | 83.7253 | 15.4829 |
| 0.0716 | 5.87 | 10400 | 0.4788 | 84.3068 | 76.7969 | 83.74 | 83.7402 | 15.4747 |
| 0.0613 | 6.09 | 10800 | 0.5252 | 84.4256 | 77.008 | 83.8688 | 83.8828 | 15.4815 |
| 0.0439 | 6.32 | 11200 | 0.5398 | 84.3753 | 76.8235 | 83.793 | 83.7986 | 15.4815 |
| 0.0452 | 6.55 | 11600 | 0.5377 | 84.4467 | 76.8923 | 83.8893 | 83.8818 | 15.4815 |
| 0.0434 | 6.77 | 12000 | 0.5347 | 84.3734 | 76.811 | 83.8108 | 83.8063 | 15.4843 |
| 0.0424 | 7.0 | 12400 | 0.5380 | 84.4558 | 76.9239 | 83.9033 | 83.9022 | 15.4751 |
| 0.0296 | 7.22 | 12800 | 0.5808 | 84.332 | 76.8729 | 83.7923 | 83.7826 | 15.4774 |
| 0.0287 | 7.45 | 13200 | 0.5956 | 84.4744 | 77.0945 | 83.9222 | 83.9228 | 15.4843 |
| 0.0283 | 7.67 | 13600 | 0.5966 | 84.4271 | 77.0661 | 83.877 | 83.8712 | 15.4829 |
| 0.0285 | 7.9 | 14000 | 0.5983 | 84.4562 | 77.0334 | 83.8987 | 83.8985 | 15.4824 |
Framework versions
- Transformers 4.28.1
- Pytorch 1.11.0a0+b6df043
- Datasets 2.12.0
- Tokenizers 0.13.3
