distilbert-base-uncased-distilled-clinc

This model is a fine-tuned version of distilbert-base-uncased on the clinc_oos dataset. It achieves the following results on the evaluation set:

• Loss: 0.2869
• Accuracy: 0.9503

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: 48
• eval_batch_size: 48
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.026785267717638298
• num_epochs: 24

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	318	2.1228	0.7194
2.5433	2.0	636	0.8036	0.8935
2.5433	3.0	954	0.4630	0.9355
0.7139	4.0	1272	0.3767	0.9429
0.3352	5.0	1590	0.3417	0.9461
0.3352	6.0	1908	0.3249	0.95
0.2555	7.0	2226	0.3141	0.9487
0.2237	8.0	2544	0.3089	0.9490
0.2237	9.0	2862	0.3039	0.9487
0.2098	10.0	3180	0.3040	0.9487
0.2098	11.0	3498	0.2971	0.9516
0.2004	12.0	3816	0.2945	0.95
0.1949	13.0	4134	0.2967	0.9468
0.1949	14.0	4452	0.2912	0.9497
0.1905	15.0	4770	0.2907	0.9513
0.1883	16.0	5088	0.2927	0.9487
0.1883	17.0	5406	0.2901	0.9503
0.1852	18.0	5724	0.2879	0.9497
0.184	19.0	6042	0.2895	0.95
0.184	20.0	6360	0.2876	0.9519
0.1828	21.0	6678	0.2871	0.9503
0.1828	22.0	6996	0.2867	0.9510
0.1816	23.0	7314	0.2868	0.9503
0.1813	24.0	7632	0.2869	0.9503

Framework versions

• Transformers 4.16.2
• Pytorch 2.4.1+cu121
• Datasets 1.16.1
• Tokenizers 0.19.1