distilbert-complaints-wandb
This model is a fine-tuned version of distilbert-base-uncased on the consumer-finance-complaints dataset. It achieves the following results on the evaluation set:
- Loss: 0.4448
- Accuracy: 0.8689
- F1: 0.8631
- Recall: 0.8689
- Precision: 0.8616
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: 5e-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
- lr_scheduler_warmup_steps: 500
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision |
---|---|---|---|---|---|---|---|
0.571 | 0.51 | 2000 | 0.5150 | 0.8469 | 0.8349 | 0.8469 | 0.8249 |
0.4765 | 1.01 | 4000 | 0.4676 | 0.8561 | 0.8451 | 0.8561 | 0.8376 |
0.3376 | 1.52 | 6000 | 0.4560 | 0.8609 | 0.8546 | 0.8609 | 0.8547 |
0.268 | 2.03 | 8000 | 0.4399 | 0.8684 | 0.8611 | 0.8684 | 0.8607 |
0.2654 | 2.53 | 10000 | 0.4448 | 0.8689 | 0.8631 | 0.8689 | 0.8616 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.11.0+cu113
- Datasets 2.3.2
- Tokenizers 0.12.1
- Downloads last month
- 2
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.
Dataset used to train Kayvane/distilbert-complaints-wandb
Evaluation results
- Accuracy on consumer-finance-complaintsself-reported0.869
- F1 on consumer-finance-complaintsself-reported0.863
- Recall on consumer-finance-complaintsself-reported0.869
- Precision on consumer-finance-complaintsself-reported0.862