distilbert-base-uncased-finetuned-ingredients
This model is a fine-tuned version of distilbert-base-uncased on the ingredients_yes_no dataset. It achieves the following results on the evaluation set:
- Loss: 0.0105
- Precision: 0.9899
- Recall: 0.9932
- F1: 0.9915
- Accuracy: 0.9978
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: 16
- eval_batch_size: 16
- seed: 42
- 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 | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
No log | 1.0 | 47 | 0.2783 | 0.4 | 0.5492 | 0.4629 | 0.8910 |
No log | 2.0 | 94 | 0.1089 | 0.8145 | 0.8780 | 0.8450 | 0.9718 |
No log | 3.0 | 141 | 0.0273 | 0.9865 | 0.9932 | 0.9899 | 0.9973 |
No log | 4.0 | 188 | 0.0168 | 0.9865 | 0.9932 | 0.9899 | 0.9973 |
No log | 5.0 | 235 | 0.0156 | 0.9865 | 0.9898 | 0.9882 | 0.9957 |
No log | 6.0 | 282 | 0.0129 | 0.9865 | 0.9932 | 0.9899 | 0.9973 |
No log | 7.0 | 329 | 0.0121 | 0.9899 | 0.9932 | 0.9915 | 0.9978 |
No log | 8.0 | 376 | 0.0115 | 0.9899 | 0.9932 | 0.9915 | 0.9978 |
No log | 9.0 | 423 | 0.0108 | 0.9899 | 0.9932 | 0.9915 | 0.9978 |
No log | 10.0 | 470 | 0.0105 | 0.9899 | 0.9932 | 0.9915 | 0.9978 |
Framework versions
- Transformers 4.10.2
- Pytorch 1.9.0+cu102
- Datasets 1.11.0
- Tokenizers 0.10.3
- Downloads last month
- 33
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Space using harr/distilbert-base-uncased-finetuned-ingredients 1
Evaluation results
- Precision on ingredients_yes_noself-reported0.990
- Recall on ingredients_yes_noself-reported0.993
- F1 on ingredients_yes_noself-reported0.992
- Accuracy on ingredients_yes_noself-reported0.998