sentiment_oversampling_25_12

This model is a fine-tuned version of VietAI/vit5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1955
• F1: 0.6240
• Accuracy: 0.7911

Model description

Sentiment analysis for user's review on Meta application

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: 1e-05
• train_batch_size: 80
• eval_batch_size: 80
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Accuracy
1.2607	0.37	200	0.2220	0.5931	0.7885
0.2748	0.75	400	0.2099	0.6012	0.7954
0.2591	1.12	600	0.1922	0.6114	0.8117
0.2519	1.5	800	0.1993	0.6203	0.7975
0.2442	1.87	1000	0.2153	0.6092	0.7555
0.2408	2.24	1200	0.1915	0.6182	0.7992
0.2309	2.62	1400	0.2064	0.6109	0.7703
0.2322	2.99	1600	0.1972	0.6169	0.7881

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.0
• Tokenizers 0.15.0