sentiment-analysis-model

This model is a fine-tuned version of dumitrescustefan/bert-base-romanian-cased-v1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2433
• Accuracy: 0.9151
• F1: 0.9151

Intended uses & limitations

This model is intended to label reviews written in romanian as either POSITIVE or NEGATIVE.

Training and evaluation data

Trained and evaluated using half of ro_sent dataset.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• 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: linear
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	281	0.2449	0.9017	0.9020
0.2392	2.0	562	0.2433	0.9151	0.9151

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2