dts_ESLO.06.05.25_exp.ft.dia.1.A_mdl.no3

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: 0.7585
• Model Preparation Time: 0.0041
• Der: 0.4608
• False Alarm: 0.1507
• Missed Detection: 0.2148
• Confusion: 0.0953

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	0.8091	0.0041	0.5159	0.1392	0.2577	0.1191
0.8156	2.0	540	0.7667	0.0041	0.4795	0.1447	0.2335	0.1013
0.7847	3.0	810	0.7597	0.0041	0.4837	0.1635	0.2073	0.1128
0.7595	4.0	1080	0.7597	0.0041	0.4760	0.1684	0.1995	0.1082
0.7789	5.0	1350	0.7471	0.0041	0.4709	0.1506	0.2250	0.0952
0.7656	6.0	1620	0.7597	0.0041	0.4730	0.1545	0.2184	0.1002
0.7307	7.0	1890	0.7614	0.0041	0.4688	0.1451	0.2308	0.0929
0.7318	8.0	2160	0.7588	0.0041	0.4690	0.1537	0.2131	0.1021
0.7115	9.0	2430	0.7555	0.0041	0.4636	0.1648	0.1932	0.1056
0.7258	10.0	2700	0.7594	0.0041	0.4743	0.1596	0.2074	0.1073
0.7144	11.0	2970	0.7679	0.0041	0.4713	0.1348	0.2435	0.0930
0.7214	12.0	3240	0.7462	0.0041	0.4617	0.1497	0.2158	0.0962
0.6578	13.0	3510	0.7601	0.0041	0.4654	0.1474	0.2219	0.0961
0.6909	14.0	3780	0.7701	0.0041	0.4620	0.1459	0.2236	0.0925
0.6775	15.0	4050	0.7600	0.0041	0.4637	0.1507	0.2154	0.0976
0.6771	16.0	4320	0.7552	0.0041	0.4607	0.1533	0.2089	0.0985
0.69	17.0	4590	0.7577	0.0041	0.4643	0.1497	0.2188	0.0958
0.6816	18.0	4860	0.7570	0.0041	0.4620	0.1503	0.2160	0.0957
0.6924	19.0	5130	0.7587	0.0041	0.4606	0.1507	0.2151	0.0948
0.7091	20.0	5400	0.7585	0.0041	0.4608	0.1507	0.2148	0.0953

Framework versions

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