<!-- 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. -->
synpre_mix_v1_1M_t5-base
This model is a fine-tuned version of t5-base on the tyzhu/synpre_mix_v1_1M dataset. It achieves the following results on the evaluation set:
- Loss: 0.0373
- Bleu: 95.1597
- Gen Len: 78.1145
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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 10000
- training_steps: 80000
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 9.3107 | 0.64 | 5000 | 9.2089 | 0.1408 | 188.7136 |
| 8.8855 | 1.28 | 10000 | 8.7474 | 0.9736 | 46.695 |
| 7.9405 | 1.92 | 15000 | 7.5952 | 1.0093 | 34.5102 |
| 0.2142 | 2.56 | 20000 | 0.0964 | 81.1212 | 87.1969 |
| 0.0486 | 3.2 | 25000 | 0.0368 | 96.4553 | 77.2838 |
| 0.0407 | 3.84 | 30000 | 0.0348 | 96.2365 | 77.3918 |
| 0.037 | 4.48 | 35000 | 0.0334 | 96.498 | 77.249 |
| 0.0374 | 5.12 | 40000 | 0.0336 | 97.203 | 76.605 |
| 0.0345 | 5.76 | 45000 | 0.0336 | 96.6274 | 77.0764 |
| 0.0369 | 6.4 | 50000 | 0.0314 | 95.9822 | 77.4934 |
| 0.0401 | 7.04 | 55000 | 0.0339 | 95.9164 | 77.4963 |
| 0.0387 | 7.68 | 60000 | 0.0343 | 94.6771 | 78.4242 |
| 0.0463 | 8.32 | 65000 | 0.0354 | 94.8001 | 78.4383 |
| 0.0471 | 8.96 | 70000 | 0.0368 | 94.3453 | 78.6396 |
| 0.0495 | 9.6 | 75000 | 0.0388 | 94.7903 | 78.0774 |
| 0.0542 | 10.24 | 80000 | 0.0373 | 95.1597 | 78.1145 |
Framework versions
- Transformers 4.34.0
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.14.1
