<!-- 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_v2_1M_t5-small
This model is a fine-tuned version of t5-small on the tyzhu/synpre_mix_v2_1M dataset. It achieves the following results on the evaluation set:
- Loss: 0.0168
- Bleu: 99.3181
- Gen Len: 87.3774
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 |
|---|---|---|---|---|---|
| 9.6556 | 0.64 | 5000 | 9.4973 | 0.1248 | 139.8121 |
| 9.2938 | 1.28 | 10000 | 9.0715 | 0.5017 | 44.2403 |
| 8.6364 | 1.92 | 15000 | 8.0280 | 0.7569 | 42.6634 |
| 2.3144 | 2.56 | 20000 | 0.4777 | 68.0223 | 98.2618 |
| 0.3528 | 3.2 | 25000 | 0.1284 | 87.8919 | 93.7593 |
| 0.1066 | 3.84 | 30000 | 0.0374 | 97.5693 | 87.9799 |
| 0.0513 | 4.48 | 35000 | 0.0317 | 97.8534 | 88.1064 |
| 0.0365 | 5.12 | 40000 | 0.0303 | 98.1907 | 87.8268 |
| 0.031 | 5.76 | 45000 | 0.0302 | 97.8241 | 88.3285 |
| 0.0262 | 6.4 | 50000 | 0.0269 | 98.0485 | 88.2178 |
| 0.0222 | 7.04 | 55000 | 0.0243 | 98.8827 | 87.5605 |
| 0.0214 | 7.68 | 60000 | 0.0229 | 98.8439 | 87.636 |
| 0.0175 | 8.32 | 65000 | 0.0215 | 98.8189 | 87.7112 |
| 0.0163 | 8.96 | 70000 | 0.0200 | 99.0476 | 87.5492 |
| 0.0145 | 9.6 | 75000 | 0.0185 | 99.255 | 87.3879 |
| 0.0152 | 10.24 | 80000 | 0.0168 | 99.3181 | 87.3774 |
Framework versions
- Transformers 4.34.0
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.14.1
