diagnostic-intent-classification-v0.0.6

This model is a fine-tuned version of MoritzLaurer/mDeBERTa-v3-base-mnli-xnli on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2907
• Accuracy: 0.9169
• F1: 0.9169
• Precision: 0.9169
• Recall: 0.9169

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 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: 500
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.7904	0.2415	100	0.8177	0.6165	0.4082	0.4165	0.4100
0.6348	0.4831	200	0.4923	0.7948	0.7947	0.7948	0.7946
0.4864	0.7246	300	0.4107	0.8393	0.8393	0.8393	0.8393
0.4435	0.9662	400	0.4024	0.8363	0.8363	0.8364	0.8362
0.4082	1.2077	500	0.3633	0.8565	0.8565	0.8568	0.8567
0.3906	1.4493	600	0.3705	0.8590	0.8588	0.8604	0.8586
0.3616	1.6908	700	0.3362	0.8702	0.8700	0.8711	0.8699
0.3589	1.9324	800	0.3398	0.8732	0.8731	0.8733	0.8731
0.3091	2.1739	900	0.3378	0.8753	0.8751	0.8769	0.8749
0.2855	2.4155	1000	0.3178	0.8877	0.8877	0.8878	0.8878
0.2847	2.6570	1100	0.3261	0.8852	0.8851	0.8857	0.8849
0.2791	2.8986	1200	0.3021	0.8959	0.8958	0.8961	0.8957
0.2169	3.1401	1300	0.3393	0.8976	0.8975	0.8979	0.8974
0.212	3.3816	1400	0.3058	0.8976	0.8976	0.8976	0.8976
0.2044	3.6232	1500	0.3098	0.9100	0.9100	0.9100	0.9100
0.2124	3.8647	1600	0.2795	0.9126	0.9126	0.9127	0.9125
0.1731	4.1063	1700	0.3008	0.9135	0.9134	0.9134	0.9134
0.1703	4.3478	1800	0.3031	0.9117	0.9117	0.9117	0.9117
0.1622	4.5894	1900	0.2925	0.9135	0.9134	0.9135	0.9134
0.1588	4.8309	2000	0.2907	0.9169	0.9169	0.9169	0.9169

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0