deberta-base-finetuned-squad-v2

This model is a fine-tuned version of microsoft/deberta-base on the squad_v2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.9221

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.753	1.0	8238	0.7286
0.5378	2.0	16476	0.7578
0.3881	3.0	24714	0.9221

Performance

• 'exact': 81.84115219405373
• 'f1': 85.19125695340612
• 'total': 11873
• 'HasAns_exact': 80.24628879892038
• 'HasAns_f1': 86.95610556811602
• 'HasAns_total': 5928
• 'NoAns_exact': 83.43145500420522
• 'NoAns_f1': 83.43145500420522
• 'NoAns_total': 5945
• 'best_exact': 81.84115219405373
• 'best_exact_thresh': 0.9994916319847107
• 'best_f1': 85.19125695340657
• 'best_f1_thresh': 0.9994916319847107
• 'total_time_in_seconds': 294.34524957099984
• 'samples_per_second': 40.33698528277447
• 'latency_in_seconds': 0.024791143735450168

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3