nli-subgroups-target-abroad

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

• Loss: 1.3895
• Accuracy: 0.9270
• Precision Binary: 0.5491
• Recall Binary: 0.6333
• F1 Binary: 0.5882
• Precision Micro: 0.9270
• Recall Micro: 0.9270
• F1 Micro: 0.9270
• Pr Auc: 0.5737
• Cohen Kappa: 0.5484

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: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision Binary	Recall Binary	F1 Binary	Precision Micro	Recall Micro	F1 Micro	Pr Auc	Cohen Kappa
No log	1.0	456	1.0509	0.9237	0.5307	0.6333	0.5775	0.9237	0.9237	0.9237	0.5397	0.5359
1.1638	2.0	912	0.7448	0.9303	0.5782	0.5667	0.5724	0.9303	0.9303	0.9303	0.5361	0.5345
0.9414	3.0	1368	1.3895	0.9270	0.5491	0.6333	0.5882	0.9270	0.9270	0.9270	0.5737	0.5484

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1