dts_ESLO.06.05.25_exp.ft.dia.1.A_mdl.no2

This model is a fine-tuned version of pyannote/segmentation-3.0 on the CAENNAIS dataset. It achieves the following results on the evaluation set:

• Loss: 0.7678
• Model Preparation Time: 0.0043
• Der: 0.4560
• False Alarm: 0.1445
• Missed Detection: 0.2165
• Confusion: 0.0950

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: 0.001
• 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: cosine
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Der	False Alarm	Missed Detection	Confusion
0.8687	1.0	270	0.8003	0.0043	0.5137	0.1447	0.2577	0.1113
0.8032	2.0	540	0.7798	0.0043	0.5002	0.1203	0.2856	0.0943
0.7846	3.0	810	0.7556	0.0043	0.4809	0.1545	0.2251	0.1013
0.7547	4.0	1080	0.7515	0.0043	0.4625	0.1559	0.2046	0.1021
0.7571	5.0	1350	0.7644	0.0043	0.4757	0.1400	0.2389	0.0968
0.7751	6.0	1620	0.7725	0.0043	0.4773	0.1317	0.2519	0.0938
0.7187	7.0	1890	0.7781	0.0043	0.4726	0.1338	0.2455	0.0933
0.7351	8.0	2160	0.7491	0.0043	0.4640	0.1362	0.2408	0.0870
0.7112	9.0	2430	0.7530	0.0043	0.4610	0.1535	0.2129	0.0946
0.7275	10.0	2700	0.7666	0.0043	0.4597	0.1377	0.2329	0.0891
0.7159	11.0	2970	0.7624	0.0043	0.4567	0.1375	0.2314	0.0878
0.7201	12.0	3240	0.7474	0.0043	0.4598	0.1410	0.2271	0.0917
0.6565	13.0	3510	0.7511	0.0043	0.4579	0.1475	0.2137	0.0966
0.6864	14.0	3780	0.7658	0.0043	0.4636	0.1429	0.2278	0.0930
0.695	15.0	4050	0.7739	0.0043	0.4557	0.1459	0.2143	0.0955
0.6775	16.0	4320	0.7636	0.0043	0.4541	0.1467	0.2139	0.0935
0.6821	17.0	4590	0.7694	0.0043	0.4565	0.1444	0.2166	0.0955
0.6742	18.0	4860	0.7674	0.0043	0.4575	0.1441	0.2181	0.0952
0.693	19.0	5130	0.7672	0.0043	0.4557	0.1441	0.2168	0.0948
0.7177	20.0	5400	0.7678	0.0043	0.4560	0.1445	0.2165	0.0950

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0