This model is a fine-tuned version of pyannote/speaker-diarization-3.0 on the Beijuka/hf_luganda_mental_health_dataset dataset. It achieves the following results on the evaluation set:

Loss: 0.2251
Model Preparation Time: 0.0171
Der: 0.0933
False Alarm: 0.0237
Missed Detection: 0.0387
Confusion: 0.0309

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: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Der	False Alarm	Missed Detection	Confusion
0.3335	1.0	114	0.2926	0.0171	0.1212	0.0378	0.0501	0.0333
0.2908	2.0	228	0.2792	0.0171	0.1137	0.0434	0.0439	0.0264
0.2660	3.0	342	0.2788	0.0171	0.1117	0.0408	0.0447	0.0262
0.2431	4.0	456	0.2868	0.0171	0.1131	0.0361	0.0509	0.0261
0.2398	5.0	570	0.2723	0.0171	0.1093	0.0354	0.0495	0.0244
0.2457	6.0	684	0.2685	0.0171	0.1067	0.0411	0.0433	0.0223
0.2368	7.0	798	0.2697	0.0171	0.1100	0.0422	0.0440	0.0239
0.2211	8.0	912	0.2699	0.0171	0.1091	0.0416	0.0444	0.0230
0.2258	9.0	1026	0.2697	0.0171	0.1089	0.0410	0.0446	0.0233

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Beijuka/speaker-segmentation-sula-hf_luganda_mental_health_dataset-v1S

speaker-segmentation-sula-hf_luganda_mental_health_dataset-v1S

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Beijuka/
speaker-segmentation-sula-hf_luganda_mental_health_dataset-v1S