<!-- 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. -->
flan-t5-large-da-multiwoz2.0_400-ep10-nonstop
This model is a fine-tuned version of google/flan-t5-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3642
- Accuracy: 40.0483
- Num: 7358
- Gen Len: 15.3939
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: 8
- eval_batch_size: 32
- seed: 1799
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Num | Gen Len |
|---|---|---|---|---|---|---|
| 1.3288 | 0.58 | 200 | 0.5718 | 27.6241 | 7358 | 13.9315 |
| 0.5941 | 1.17 | 400 | 0.4745 | 32.0484 | 7358 | 14.2517 |
| 0.5213 | 1.75 | 600 | 0.4283 | 33.6716 | 7358 | 15.1218 |
| 0.4891 | 2.33 | 800 | 0.4063 | 34.1115 | 7358 | 14.8005 |
| 0.4503 | 2.92 | 1000 | 0.3935 | 35.9067 | 7358 | 15.9095 |
| 0.4213 | 3.5 | 1200 | 0.3833 | 37.2591 | 7358 | 16.0005 |
| 0.4184 | 4.08 | 1400 | 0.3795 | 37.696 | 7358 | 15.478 |
| 0.3959 | 4.66 | 1600 | 0.3762 | 37.401 | 7358 | 14.8752 |
| 0.3847 | 5.25 | 1800 | 0.3714 | 37.7347 | 7358 | 15.9304 |
| 0.3779 | 5.83 | 2000 | 0.3710 | 38.6814 | 7358 | 14.9257 |
| 0.3776 | 6.41 | 2200 | 0.3681 | 38.4266 | 7358 | 15.2517 |
| 0.3601 | 7.0 | 2400 | 0.3669 | 38.6749 | 7358 | 15.1791 |
| 0.3504 | 7.58 | 2600 | 0.3669 | 39.2748 | 7358 | 15.4308 |
| 0.3568 | 8.16 | 2800 | 0.3650 | 39.798 | 7358 | 15.8966 |
| 0.3528 | 8.75 | 3000 | 0.3630 | 39.912 | 7358 | 15.4081 |
| 0.3463 | 9.33 | 3200 | 0.3641 | 40.1243 | 7358 | 15.5367 |
| 0.3439 | 9.91 | 3400 | 0.3642 | 40.0567 | 7358 | 15.3842 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.5.1
- Tokenizers 0.12.1
