finetuning-pysentimiento-war-tweets

This model is a fine-tuned version of finiteautomata/beto-sentiment-analysis on a dataset of 1500 tweets from Peruvian accounts. It achieves the following results on the evaluation set:

• Loss: 1.7689
• Accuracy: 0.7378
• F1: 0.7456

Model description

This model in a fine-tuned version of finiteautomata/beto-sentiment-analysis using five labels: pro_russia, against_ukraine, neutral, against_russia, pro_ukraine.

Intended uses & limitations

This model shall be used to classify text (more specifically, Spanish tweets) as expressing a position concerning the Russo-Ukrainian war.

Training and evaluation data

We used an 80/20 training/test split on the aforementioned dataset.

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: 30

Framework versions

• Transformers 4.20.1
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1