deberta-financial

This model is a fine-tuned version of microsoft/deberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2137
• Accuracy: 0.9146
• Filtered Samples: 0
• Remaining Samples: 2751

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Filtered Samples	Remaining Samples
0.4647	0.2997	232	0.3944	0.8593	0	2751
0.3971	0.5995	464	0.3618	0.8582	0	2751
0.2951	0.8992	696	0.3034	0.8917	0	2751
0.2265	1.1990	928	0.2843	0.8964	0	2751
0.269	1.4987	1160	0.2540	0.9059	0	2751
0.2077	1.7984	1392	0.2481	0.9037	0	2751
0.1427	2.0982	1624	0.2283	0.9073	0	2751
0.1723	2.3979	1856	0.2224	0.9124	0	2751
0.1759	2.6977	2088	0.2198	0.9131	0	2751
0.1836	2.9974	2320	0.2137	0.9146	0	2751

Framework versions

• Transformers 4.51.3
• Pytorch 2.4.0a0+git7cecbf6
• Datasets 3.5.0
• Tokenizers 0.21.1