dougtrajano/toxic-comment-classification

This model is a fine-tuned version of neuralmind/bert-large-portuguese-cased on the OLID-BR dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4102
  • Accuracy: 0.8547
  • F1: 0.8549
  • Precision: 0.8669
  • Recall: 0.8547

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: 3.255788747459486e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 1993
  • optimizer: Adam with betas=(0.8445637934160373,0.8338816842140165) and epsilon=2.527092625455385e-08
  • lr_scheduler_type: linear
  • num_epochs: 30
  • label_smoothing_factor: 0.07158711257743958

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Precision Recall
0.4465 1.0 1408 0.4102 0.8547 0.8549 0.8669 0.8547
0.3839 2.0 2816 0.4814 0.8509 0.8497 0.8532 0.8509
0.3945 3.0 4224 0.6362 0.8002 0.7918 0.8258 0.8002
0.3643 4.0 5632 0.4961 0.8248 0.8211 0.8349 0.8248
0.3345 5.0 7040 0.5267 0.8528 0.8532 0.8570 0.8528
0.3053 6.0 8448 0.5902 0.8002 0.7911 0.8292 0.8002

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.10.2+cu113
  • Datasets 2.9.0
  • Tokenizers 0.13.2
Downloads last month
1,344
Safetensors
Model size
334M params
Tensor type
I64
·
F32
·
Inference Examples
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.

Model tree for dougtrajano/toxic-comment-classification

Finetuned
(34)
this model