finetuning-sentiment-text-mining

• Loss: 1.3600
• Accuracy: 0.784
• F1: 0.7840

Model description

This model was finetuned for course Text mining for Ai at Vrije university Amsterdam

Intended uses & limitations

Only use is for the final project in the course

Training and evaluation data

Data were reddit and twitter comments or tweets

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: 10

Training results

Framework versions

• Transformers 4.27.3
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2