deberta-v3-ielts-node-classifier-v3

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: 1.3429
• Accuracy: 0.5085
• F1 Macro: 0.1685

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 OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 6
• label_smoothing_factor: 0.05

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
3.6869	0.6369	100	3.4232	0.1955	0.0818
1.4881	1.2739	200	1.5163	0.1966	0.0822
1.4256	1.9108	300	1.3756	0.1966	0.0822
1.3842	2.5478	400	1.3779	0.5085	0.1685
1.3608	3.1847	500	1.3413	0.5085	0.1685
1.3480	3.8217	600	1.3528	0.5085	0.1685
1.3431	4.4586	700	1.3453	0.5085	0.1685
1.3597	5.0955	800	1.3448	0.5085	0.1685
1.3451	5.7325	900	1.3430	0.5085	0.1685
1.3451	6.0	942	1.3429	0.5085	0.1685

Framework versions

• Transformers 5.5.4
• Pytorch 2.10.0+cu130
• Datasets 3.0.0
• Tokenizers 0.22.2