deberta-qqp

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

• Loss: 0.2173
• Accuracy: 0.9215

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: 64
• eval_batch_size: 512
• 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
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
0.2847	0.1759	1000	0.8607	0.3087
0.2686	0.3517	2000	0.8884	0.2690
0.2302	0.5276	3000	0.8978	0.2448
0.2528	0.7035	4000	0.9042	0.2267
0.2394	0.8794	5000	0.9078	0.2212
0.1934	1.0552	6000	0.9113	0.2257
0.1723	1.2311	7000	0.9137	0.2237
0.1883	1.4070	8000	0.9154	0.2071
0.1775	1.5828	9000	0.9155	0.2082
0.1823	1.7587	10000	0.9188	0.2031
0.1548	1.9346	11000	0.2162	0.9184
0.1289	2.1104	12000	0.2311	0.9177
0.1309	2.2863	13000	0.2231	0.9187
0.1252	2.4622	14000	0.2182	0.9195
0.1306	2.6381	15000	0.2099	0.9204
0.1344	2.8139	16000	0.2152	0.9212
0.1231	2.9898	17000	0.2173	0.9215

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.1+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2