AnnoMI-full_speaker_role_id-bert-base-uncased-v1

• Loss: 0.3878
• Accuracy: 0.8819
• Precision Macro: 0.8819
• Recall Macro: 0.8819
• F1 Macro: 0.8819
• Precision Weighted: 0.8820
• Recall Weighted: 0.8819
• F1 Weighted: 0.8819
• Report: precision recall f1-score support

       0       0.89      0.88      0.88       683
       1       0.88      0.89      0.88       672

  accuracy 0.88 1355 macro avg 0.88 0.88 0.88 1355 weighted avg 0.88 0.88 0.88 1355

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

       0       0.90      0.87      0.89       683
       1       0.87      0.90      0.89       672

accuracy                           0.89      1355

       0       0.90      0.86      0.88       683
       1       0.87      0.90      0.88       672

accuracy                           0.88      1355

       0       0.89      0.87      0.88       683
       1       0.87      0.89      0.88       672

accuracy                           0.88      1355

       0       0.89      0.88      0.88       683
       1       0.88      0.89      0.88       672

accuracy                           0.88      1355

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.2