tiny-bert-Sentiment-persian

This model is a fine-tuned version of dadashzadeh/tiny-bert-Sentiment-persian on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6553
• Accuracy: 0.7611

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: 2
• eval_batch_size: 2
• seed: 45
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6157	0.9999	3575	0.6703	0.7577
0.5833	1.9999	7150	0.7599	0.7171
0.6015	2.9998	10725	0.6824	0.7590
0.5601	4.0	14301	0.6780	0.7533
0.5699	4.9999	17876	0.7071	0.7356
0.5519	5.9999	21451	0.6931	0.7391
0.5436	6.9998	25026	0.6736	0.7629
0.5482	8.0	28602	0.6567	0.7685
0.5367	8.9999	32177	0.6553	0.7611
0.5399	9.9999	35752	0.6691	0.7616
0.5112	10.9998	39327	0.6785	0.7564
0.5113	11.9992	42900	0.6773	0.7572

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.2+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1