userutterance_classification_ver1

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

• Loss: 0.2898
• Accuracy: 0.9539

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: 4e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 130
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.8334	0.15	200	4.7254	0.0748
3.4798	0.3	400	3.4244	0.2971
2.319	0.45	600	2.4423	0.5184
1.5683	0.6	800	1.7401	0.6310
0.9625	0.75	1000	1.2750	0.7265
0.6922	0.9	1200	0.9717	0.7761
0.5019	1.05	1400	0.8036	0.8284
0.3538	1.2	1600	0.6690	0.8471
0.2413	1.35	1800	0.5585	0.8713
0.2623	1.5	2000	0.4840	0.8874
0.2103	1.66	2200	0.4261	0.9126
0.1456	1.81	2400	0.3872	0.9152
0.1276	1.96	2600	0.3329	0.9290
0.09	2.11	2800	0.2925	0.9432
0.0534	2.26	3000	0.2996	0.9361
0.0588	2.41	3200	0.2951	0.9403
0.044	2.56	3400	0.3324	0.9403
0.0535	2.71	3600	0.3155	0.9432
0.0537	2.86	3800	0.3206	0.9419
0.1325	3.01	4000	0.2945	0.9465
0.0611	3.16	4200	0.2903	0.9442
0.0077	3.31	4400	0.3052	0.9477
0.0187	3.46	4600	0.2774	0.95
0.0125	3.61	4800	0.2851	0.9513
0.0157	3.76	5000	0.2883	0.9523
0.0414	3.91	5200	0.3163	0.9497
0.0025	4.06	5400	0.2998	0.9494
0.0019	4.21	5600	0.2925	0.9513
0.0013	4.36	5800	0.2872	0.9526
0.0014	4.51	6000	0.2906	0.9532
0.0015	4.67	6200	0.2862	0.9529
0.0281	4.82	6400	0.2863	0.9535
0.0287	4.97	6600	0.2898	0.9539

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2