wcBERTaAttSpmm-ggTranslate-senticPolarEmotion-bully-f2

This model is a fine-tuned version of airesearch/wangchanberta-base-att-spm-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5206
• Accuracy: 0.7354
• Precision: 0.7320
• Recall: 0.7354
• F1 Score: 0.7335

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: 5e-06
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 Score
0.6386	1.0	120	0.6121	0.7008	0.4912	0.7008	0.5776
0.603	2.0	240	0.5663	0.7144	0.6835	0.7144	0.6762
0.5758	3.0	360	0.5828	0.7029	0.6989	0.7029	0.7007
0.5724	4.0	480	0.5518	0.7238	0.6978	0.7238	0.6894
0.5412	5.0	600	0.5336	0.7082	0.6926	0.7082	0.6975
0.5295	6.0	720	0.5222	0.7280	0.7111	0.7280	0.7145
0.4851	7.0	840	0.5090	0.7312	0.7136	0.7312	0.7162
0.4639	8.0	960	0.5153	0.7259	0.7164	0.7259	0.7200
0.4606	9.0	1080	0.5233	0.7029	0.7241	0.7029	0.7104
0.4412	10.0	1200	0.5501	0.6998	0.7496	0.6998	0.7116
0.4106	11.0	1320	0.5262	0.7155	0.7398	0.7155	0.7235
0.3976	12.0	1440	0.5148	0.7374	0.7300	0.7374	0.7329
0.3917	13.0	1560	0.5194	0.7364	0.7338	0.7364	0.7350
0.3858	14.0	1680	0.5225	0.7259	0.7293	0.7259	0.7275
0.3853	15.0	1800	0.5206	0.7354	0.7320	0.7354	0.7335

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1