roberta-base-finetuned-intent-ipu
This model is a fine-tuned version of roberta-base on the snips_built_in_intents dataset. It achieves the following results on the evaluation set:
- Loss: 0.1503
- Accuracy: 1.0
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: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: IPU
- gradient_accumulation_steps: 8
- total_train_batch_size: 8
- total_eval_batch_size: 5
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- training precision: Mixed Precision
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.2478 | 1.0 | 75 | 0.6069 | 0.96 |
| 0.2522 | 2.0 | 150 | 0.1503 | 1.0 |
| 0.0903 | 3.0 | 225 | 0.0712 | 1.0 |
| 0.0883 | 4.0 | 300 | 0.0350 | 1.0 |
| 0.0491 | 5.0 | 375 | 0.0267 | 1.0 |
| 0.0305 | 6.0 | 450 | 0.0218 | 1.0 |
| 0.0461 | 7.0 | 525 | 0.0191 | 1.0 |
| 0.039 | 8.0 | 600 | 0.0174 | 1.0 |
| 0.0337 | 9.0 | 675 | 0.0166 | 1.0 |
| 0.0164 | 10.0 | 750 | 0.0162 | 1.0 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.10.0+cpu
- Datasets 2.7.1
- Tokenizers 0.12.0