deberta-v3-xsmall-beavertails-harmful-qa-classifier

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

• Loss: 0.3043
• Accuracy: 0.8716
• Macro F1: 0.8716
• Macro Precision: 0.8738
• Macro Recall: 0.8736
• Micro F1: 0.8716
• Micro Precision: 0.8716
• Micro Recall: 0.8716

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: 6e-06
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro F1	Macro Precision	Macro Recall	Micro F1	Micro Precision	Micro Recall
0.3726	1.0	2349	0.3409	0.8573	0.8573	0.8581	0.8587	0.8573	0.8573	0.8573
0.3299	2.0	4698	0.3186	0.8669	0.8669	0.8697	0.8692	0.8669	0.8669	0.8669
0.3181	3.0	7047	0.3092	0.8683	0.8682	0.8715	0.8707	0.8683	0.8683	0.8683
0.3064	4.0	9396	0.3043	0.8716	0.8716	0.8738	0.8736	0.8716	0.8716	0.8716

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1