deberta-base-clickbait-task1-20-epoch-post_title

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

• Loss: 2.5412
• Accuracy: 0.7025

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
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	200	0.7414	0.7075
No log	2.0	400	0.6972	0.73
0.7437	3.0	600	0.7188	0.73
0.7437	4.0	800	0.9260	0.7225
0.3148	5.0	1000	1.0694	0.715
0.3148	6.0	1200	1.3980	0.735
0.3148	7.0	1400	1.6897	0.7125
0.103	8.0	1600	1.8628	0.7275
0.103	9.0	1800	2.0991	0.7125
0.0456	10.0	2000	2.0466	0.7225
0.0456	11.0	2200	2.2220	0.7225
0.0456	12.0	2400	2.3278	0.6975
0.0222	13.0	2600	2.4275	0.7025
0.0222	14.0	2800	2.4249	0.695
0.0092	15.0	3000	2.4740	0.7275
0.0092	16.0	3200	2.4897	0.7125
0.0092	17.0	3400	2.5280	0.705
0.0058	18.0	3600	2.5360	0.705
0.0058	19.0	3800	2.5075	0.715
0.0039	20.0	4000	2.5412	0.7025

Framework versions

• Transformers 4.44.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1