Edit model card

distilbert-complaints-wandb-product

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: 0.4431
  • Accuracy: 0.8691
  • F1: 0.8645
  • Recall: 0.8691
  • Precision: 0.8629

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
0.562 0.51 2000 0.5107 0.8452 0.8346 0.8452 0.8252
0.4548 1.01 4000 0.4628 0.8565 0.8481 0.8565 0.8466
0.3439 1.52 6000 0.4519 0.8605 0.8544 0.8605 0.8545
0.2626 2.03 8000 0.4412 0.8678 0.8618 0.8678 0.8626
0.2717 2.53 10000 0.4431 0.8691 0.8645 0.8691 0.8629

Framework versions

  • Transformers 4.20.1
  • Pytorch 1.11.0+cu113
  • Datasets 2.3.2
  • Tokenizers 0.12.1
Downloads last month
13

Dataset used to train Kayvane/distilbert-complaints-wandb-product

Evaluation results