mdeberta-fiqhqa-qa

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

• Loss: 0.8996
• Exact Match: 24.4604
• F1: 53.1508

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: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1
1.0956	1.0	333	0.8887	21.5827	50.0958
0.7869	2.0	666	0.8384	20.8633	53.4146
0.6457	3.0	999	0.8996	24.4604	53.1508

Framework versions

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