improved_stance_detection_v2-fold2

This model is a fine-tuned version of aubmindlab/bert-large-arabertv02-twitter on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2078
• Accuracy: 0.8098
• Macro F1: 0.8108
• Weighted F1: 0.8124
• F1 Pro: 0.8372
• F1 Against: 0.8261
• F1 Neutral: 0.7692

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 15
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro F1	Weighted F1	F1 Pro	F1 Against	F1 Neutral
2.0893	1.9320	50	0.4121	0.5073	0.4822	0.4776	0.5714	0.3093	0.5658
1.5103	3.8544	100	0.3107	0.6878	0.6886	0.6910	0.7438	0.7006	0.6212
1.0102	5.7767	150	0.2688	0.7220	0.7252	0.7273	0.7692	0.7397	0.6667
0.8139	7.6990	200	0.2152	0.7951	0.7965	0.7976	0.8254	0.7971	0.7671
0.6407	9.6214	250	0.2380	0.7951	0.7972	0.7980	0.8293	0.7910	0.7712
0.5584	11.5437	300	0.2076	0.8098	0.8108	0.8124	0.8372	0.8261	0.7692
0.4945	13.4660	350	0.2123	0.7951	0.7969	0.7986	0.8413	0.8028	0.7465

Framework versions

• Transformers 5.0.0
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2