summarization_t5base_en_to_kjven
This model is a fine-tuned version of t5-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.8324
- Bleu: 21.2143
- Gen Len: 18.1685
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: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
---|---|---|---|---|---|
1.0735 | 1.0 | 2860 | 0.9479 | 21.3913 | 18.1219 |
0.9776 | 2.0 | 5720 | 0.8750 | 22.1711 | 18.1307 |
0.918 | 3.0 | 8580 | 0.8317 | 22.6915 | 18.1381 |
0.8741 | 4.0 | 11440 | 0.8039 | 23.0856 | 18.1468 |
0.8489 | 5.0 | 14300 | 0.7841 | 23.3573 | 18.1455 |
0.8169 | 6.0 | 17160 | 0.7664 | 23.5073 | 18.1493 |
0.7965 | 7.0 | 20020 | 0.7532 | 23.6919 | 18.1495 |
0.78 | 8.0 | 22880 | 0.7411 | 23.8445 | 18.1461 |
0.7568 | 9.0 | 25740 | 0.7338 | 23.86 | 18.155 |
0.7496 | 10.0 | 28600 | 0.7228 | 23.953 | 18.1511 |
0.7411 | 11.0 | 31460 | 0.7175 | 24.0327 | 18.1511 |
0.8376 | 12.0 | 34320 | 0.8114 | 23.311 | 18.1319 |
1.1918 | 13.0 | 37180 | 0.9686 | 21.5339 | 18.1185 |
1.0929 | 14.0 | 40040 | 0.8978 | 21.561 | 18.1455 |
1.0373 | 15.0 | 42900 | 0.8617 | 21.4942 | 18.1542 |
1.0165 | 16.0 | 45760 | 0.8432 | 21.3962 | 18.1595 |
0.9973 | 17.0 | 48620 | 0.8340 | 21.2558 | 18.166 |
0.9889 | 18.0 | 51480 | 0.8326 | 21.2238 | 18.1687 |
0.9909 | 19.0 | 54340 | 0.8325 | 21.2216 | 18.1688 |
0.9942 | 20.0 | 57200 | 0.8324 | 21.2143 | 18.1685 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.