mdeberta-squad-id

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the squad_id dataset. It achieves the following results on the evaluation set:

• Loss: 0.9606
• Exact Match: 55.2800
• F1: 72.2117

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: 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: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1
1.113	1.0	8486	0.9985	52.7665	70.0577
0.8977	2.0	16972	0.9466	54.3522	71.4003
0.7574	3.0	25458	0.9606	55.2800	72.2117

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.1+cu128
• Datasets 4.4.1
• Tokenizers 0.22.1