STS-Lora-Fine-Tuning-Capstone-Deberta-old-model-pipe-test_augmentation

This model is a fine-tuned version of microsoft/deberta-v3-xsmall on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2431
• Accuracy: 0.4627

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	360	1.7500	0.2429
1.7474	2.0	720	1.7255	0.2451
1.681	3.0	1080	1.6291	0.3332
1.681	4.0	1440	1.4764	0.4130
1.5419	5.0	1800	1.4165	0.4159
1.4014	6.0	2160	1.3548	0.4336
1.3269	7.0	2520	1.3122	0.4456
1.3269	8.0	2880	1.3003	0.4529
1.2821	9.0	3240	1.2830	0.4572
1.2516	10.0	3600	1.2757	0.4576
1.2516	11.0	3960	1.2619	0.4590
1.2304	12.0	4320	1.2501	0.4670
1.2172	13.0	4680	1.2674	0.4583
1.2043	14.0	5040	1.2459	0.4656
1.2043	15.0	5400	1.2464	0.4627
1.1956	16.0	5760	1.2439	0.4645
1.1814	17.0	6120	1.2395	0.4648
1.1814	18.0	6480	1.2429	0.4637
1.1816	19.0	6840	1.2450	0.4634
1.1794	20.0	7200	1.2431	0.4627

Framework versions

• PEFT 0.10.0
• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2