<!-- 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. -->
XLM-V (base) fine-tuned on XNLI
This model is a fine-tuned version of XLM-V (base) on the XNLI (XGLUE) dataset. It achieves the following results on the evaluation set:
- Loss: 0.6511
- Accuracy: 0.7403
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: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.0994 | 0.08 | 1000 | 1.0966 | 0.3697 |
| 1.0221 | 0.16 | 2000 | 1.0765 | 0.4560 |
| 0.8437 | 0.24 | 3000 | 0.8472 | 0.6179 |
| 0.6997 | 0.33 | 4000 | 0.7650 | 0.6804 |
| 0.6304 | 0.41 | 5000 | 0.7227 | 0.7007 |
| 0.5972 | 0.49 | 6000 | 0.7430 | 0.6977 |
| 0.5886 | 0.57 | 7000 | 0.7365 | 0.7066 |
| 0.5585 | 0.65 | 8000 | 0.6819 | 0.7223 |
| 0.5464 | 0.73 | 9000 | 0.7222 | 0.7046 |
| 0.5289 | 0.81 | 10000 | 0.7290 | 0.7054 |
| 0.5298 | 0.9 | 11000 | 0.6824 | 0.7221 |
| 0.5241 | 0.98 | 12000 | 0.6650 | 0.7268 |
| 0.4806 | 1.06 | 13000 | 0.6861 | 0.7308 |
| 0.4715 | 1.14 | 14000 | 0.6619 | 0.7304 |
| 0.4645 | 1.22 | 15000 | 0.6656 | 0.7284 |
| 0.4443 | 1.3 | 16000 | 0.7026 | 0.7270 |
| 0.4582 | 1.39 | 17000 | 0.7055 | 0.7225 |
| 0.4456 | 1.47 | 18000 | 0.6592 | 0.7361 |
| 0.44 | 1.55 | 19000 | 0.6816 | 0.7329 |
| 0.4419 | 1.63 | 20000 | 0.6772 | 0.7357 |
| 0.4403 | 1.71 | 21000 | 0.6745 | 0.7319 |
| 0.4348 | 1.79 | 22000 | 0.6678 | 0.7338 |
| 0.4355 | 1.87 | 23000 | 0.6614 | 0.7365 |
| 0.4295 | 1.96 | 24000 | 0.6511 | 0.7403 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2
