amity-diarization-v00

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

• Loss: 0.3132
• Model Preparation Time: 0.0075
• Der: 0.1175
• False Alarm: 0.0395
• Missed Detection: 0.0579
• Confusion: 0.0201

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: 8.5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Der	False Alarm	Missed Detection	Confusion
No log	1.0	18	0.3317	0.0075	0.1263	0.0361	0.0699	0.0203
0.4825	2.0	36	0.3305	0.0075	0.1249	0.0366	0.0672	0.0211
0.475	3.0	54	0.3238	0.0075	0.1231	0.0369	0.0643	0.0218
0.4054	4.0	72	0.3232	0.0075	0.1213	0.0363	0.0631	0.0218
0.432	5.0	90	0.3246	0.0075	0.1206	0.0367	0.0616	0.0223
0.4053	6.0	108	0.3209	0.0075	0.1196	0.0376	0.0599	0.0222
0.4023	7.0	126	0.3193	0.0075	0.1196	0.0384	0.0589	0.0223
0.407	8.0	144	0.3195	0.0075	0.1197	0.0388	0.0585	0.0224
0.3969	9.0	162	0.3190	0.0075	0.1195	0.0388	0.0586	0.0220
0.3645	10.0	180	0.3184	0.0075	0.1197	0.0392	0.0583	0.0222
0.3645	11.0	198	0.3155	0.0075	0.1176	0.0397	0.0579	0.0200
0.3733	12.0	216	0.3223	0.0075	0.1195	0.0397	0.0578	0.0220
0.3842	13.0	234	0.3236	0.0075	0.1194	0.0397	0.0575	0.0222
0.377	14.0	252	0.3143	0.0075	0.1177	0.0396	0.0577	0.0204
0.3655	15.0	270	0.3159	0.0075	0.1176	0.0396	0.0577	0.0203
0.352	16.0	288	0.3162	0.0075	0.1176	0.0395	0.0577	0.0204
0.3753	17.0	306	0.3157	0.0075	0.1177	0.0396	0.0579	0.0201
0.3652	18.0	324	0.3169	0.0075	0.1194	0.0395	0.0579	0.0221
0.3736	19.0	342	0.3131	0.0075	0.1174	0.0394	0.0579	0.0200
0.3673	20.0	360	0.3132	0.0075	0.1175	0.0395	0.0579	0.0201

Framework versions

• Transformers 4.51.2
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1