output1

This model is a fine-tuned version of microsoft/deberta-v2-xxlarge on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7690
• Accuracy: 0.676
• Macro F1: 0.6761

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro F1
1.5278	0.2286	100	1.1249	0.5146	0.4600
0.9452	0.4571	200	0.8437	0.645	0.6425
0.8367	0.6857	300	0.8038	0.6477	0.6531
0.8092	0.9143	400	0.7801	0.6593	0.6611
0.7679	1.1429	500	0.7868	0.6717	0.6697
0.7451	1.3714	600	0.7711	0.6647	0.6645
0.7467	1.6	700	0.7646	0.6659	0.6649
0.7261	1.8286	800	0.7840	0.6649	0.6632
0.7305	2.0571	900	0.7755	0.6681	0.6707
0.6742	2.2857	1000	0.7719	0.6691	0.6707
0.6728	2.5143	1100	0.7640	0.6726	0.6726
0.6691	2.7429	1200	0.7759	0.6761	0.6783
0.677	2.9714	1300	0.7690	0.676	0.6761

Framework versions

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