<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
AIKIA_05_OFFENSIVE_BERT_MULTILINGUAL
This model is a fine-tuned version of bert-base-multilingual-uncased on the AIKIA Greek dataset with 0.5 threshold for Offensive texts. It achieves the following results on the evaluation set:
- Loss: 0.9522
- Macro F1: 0.7026
- Micro F1: 0.7085
- Accuracy: 0.7085]
Results on the test set:
-
Accuracy: 0.742483660130719
-
F1 score: 0.7400512980617473
-
Precision: 0.7458428557407848
-
Recall : 0.7399913771864991
-
Matthews Correlation Coefficient: 0.4857989934970751
-
Precision of each class: [0.72460497 0.76708075]
-
Recall of each class: [0.81060606 0.66937669]
-
F1 score of each class: [0.76519666 0.71490593]
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
- lr_scheduler_warmup_steps: 500
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Macro F1 | Micro F1 | Accuracy |
|---|---|---|---|---|---|---|
| No log | 1.0 | 338 | 0.6553 | 0.6066 | 0.6384 | 0.6384 |
| 0.6185 | 2.0 | 676 | 0.5828 | 0.7171 | 0.7232 | 0.7232 |
| 0.4438 | 3.0 | 1014 | 0.7040 | 0.7038 | 0.7085 | 0.7085 |
| 0.4438 | 4.0 | 1352 | 0.9522 | 0.7026 | 0.7085 | 0.7085 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
