metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: clickbait_binary_detection_DeBERTa
    results: []
datasets:
  - christinacdl/clickbait_notclickbait_dataset
language:
  - en
pipeline_tag: text-classification

clickbait_binary_detection_DeBERTa

This model is a fine-tuned version of microsoft/deberta-v3-large on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.7269
Macro F1: 0.9010
Micro F1: 0.9069
Accuracy: 0.9069

Performance on test set:

Accuracy: 0.911986301369863
F1 score: 0.9053903329555788
Precision: 0.9069346899004087
Recall : 0.9039394560612273
Matthews Correlation Coefficient: 0.8108686139956713
Precision of each class: [0.92560647 0.88826291]
Recall of each class: [0.93518519 0.87269373]
F1 score of each class: [0.93037117 0.88040949]

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-06
train_batch_size: 4
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Macro F1	Micro F1	Accuracy
0.2692	1.0	5475	0.2676	0.9051	0.9142	0.9142
0.2492	2.0	10951	0.3331	0.9078	0.9156	0.9156
0.2189	3.0	16426	0.3909	0.9107	0.9169	0.9169
0.1769	4.0	21902	0.3799	0.9114	0.9178	0.9178
0.1479	5.0	27377	0.5103	0.8980	0.9032	0.9032
0.108	6.0	32853	0.5215	0.9123	0.9183	0.9183
0.0957	7.0	38328	0.6549	0.8974	0.9028	0.9028
0.0773	8.0	43804	0.6768	0.9044	0.9101	0.9101
0.0586	9.0	49279	0.6837	0.9023	0.9083	0.9083
0.0439	10.0	54750	0.7269	0.9010	0.9069	0.9069

Framework versions

Transformers 4.27.1
Pytorch 2.0.1+cu118
Datasets 2.9.0
Tokenizers 0.13.3