==================================================
TRAINING CONFIGURATION
==================================================
GPU: NVIDIA RTX 6000 Ada Generation
Host IP: 132.72.180.129
Model: dicta-il/neodictabert
Output dir: /home/avramit/classifier/04122025/classifier/binary/neodictabert-finetuned-binary-041225
Batch size: 2
Learning rate: 2e-05
Epochs: 1.0
Max seq length: 4096
Labels: ['contradiction', 'entailment']
==================================================

✅ Arguments saved to /home/avramit/classifier/04122025/classifier/binary/neodictabert-finetuned-binary-041225/training_info_args.json

Dataset sizes - Train: 18238, Val: 2239, Test: 2918
Train file: /home/avramit/classifier/04122025/datasets/classifier/mergearticlesummarylevel/train.csv
Val file: /home/avramit/classifier/01122025/datasets/classifiers/validation.csv
Test file: /home/avramit/classifier/01122025/datasets/classifiers/test.csv

✅ Predictions saved to /home/avramit/classifier/04122025/classifier/binary/neodictabert-finetuned-binary-041225/predictions.txt
✅ Predictions CSV saved to /home/avramit/classifier/04122025/classifier/binary/neodictabert-finetuned-binary-041225/predictions.csv