This model is a fine-tuned version of pyannote/segmentation-3.0 on the Suraj0599/Radio_Audio1 dataset. It achieves the following results on the evaluation set:

Loss: 0.2799
Der: 0.0247
False Alarm: 0.0078
Missed Detection: 0.0043
Confusion: 0.0126

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: 64
eval_batch_size: 64
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
num_epochs: 100.0

Training results

Training Loss	Epoch	Step	Validation Loss	Der	False Alarm	Missed Detection	Confusion
0.1187	1.0	225	0.0986	0.0238	0.0077	0.0007	0.0154
0.102	2.0	450	0.0933	0.0227	0.0076	0.0008	0.0142
0.093	3.0	675	0.0917	0.0221	0.0075	0.0010	0.0136
0.0837	4.0	900	0.0867	0.0221	0.0075	0.0015	0.0130
0.0738	5.0	1125	0.0963	0.0224	0.0074	0.0015	0.0135
0.0697	6.0	1350	0.0884	0.0215	0.0075	0.0011	0.0129
0.06	7.0	1575	0.0969	0.0215	0.0074	0.0016	0.0125
0.0587	8.0	1800	0.0999	0.0209	0.0074	0.0013	0.0122
0.0536	9.0	2025	0.1006	0.0219	0.0072	0.0019	0.0128
0.0502	10.0	2250	0.1045	0.0214	0.0073	0.0017	0.0123
0.0442	11.0	2475	0.1016	0.0235	0.0072	0.0029	0.0134
0.0453	12.0	2700	0.1100	0.0226	0.0072	0.0018	0.0136
0.044	13.0	2925	0.1049	0.0211	0.0072	0.0019	0.0121
0.04	14.0	3150	0.1109	0.0210	0.0071	0.0024	0.0115
0.0441	15.0	3375	0.1142	0.0226	0.0072	0.0029	0.0125
0.0381	16.0	3600	0.1178	0.0227	0.0073	0.0038	0.0117
0.0387	17.0	3825	0.1236	0.0208	0.0071	0.0026	0.0112
0.0319	18.0	4050	0.1198	0.0228	0.0075	0.0032	0.0122
0.033	19.0	4275	0.1126	0.0222	0.0073	0.0033	0.0116
0.0345	20.0	4500	0.1233	0.0225	0.0074	0.0037	0.0115
0.0321	21.0	4725	0.1252	0.0227	0.0076	0.0037	0.0114
0.0294	22.0	4950	0.1279	0.0230	0.0075	0.0038	0.0117
0.027	23.0	5175	0.1289	0.0232	0.0075	0.0040	0.0116
0.0281	24.0	5400	0.1289	0.0223	0.0078	0.0030	0.0115
0.0269	25.0	5625	0.1308	0.0228	0.0076	0.0042	0.0111
0.027	26.0	5850	0.1339	0.0231	0.0079	0.0025	0.0127
0.0261	27.0	6075	0.1336	0.0227	0.0072	0.0040	0.0114
0.025	28.0	6300	0.1493	0.0238	0.0077	0.0029	0.0131
0.0237	29.0	6525	0.1379	0.0231	0.0077	0.0034	0.0120
0.0225	30.0	6750	0.1535	0.0227	0.0074	0.0037	0.0117
0.0211	31.0	6975	0.1472	0.0228	0.0079	0.0035	0.0114
0.0225	32.0	7200	0.1499	0.0239	0.0077	0.0040	0.0123
0.0218	33.0	7425	0.1500	0.0246	0.0076	0.0051	0.0119
0.0205	34.0	7650	0.1574	0.0237	0.0076	0.0041	0.0120
0.0216	35.0	7875	0.1497	0.0224	0.0076	0.0041	0.0107
0.0189	36.0	8100	0.1712	0.0235	0.0076	0.0047	0.0113
0.0176	37.0	8325	0.1653	0.0219	0.0076	0.0037	0.0105
0.0189	38.0	8550	0.1599	0.0225	0.0071	0.0042	0.0112
0.0172	39.0	8775	0.1835	0.0238	0.0078	0.0032	0.0128
0.0185	40.0	9000	0.2006	0.0228	0.0074	0.0027	0.0127
0.0173	41.0	9225	0.1776	0.0229	0.0077	0.0039	0.0113
0.0159	42.0	9450	0.1728	0.0239	0.0076	0.0050	0.0112
0.0166	43.0	9675	0.1831	0.0238	0.0074	0.0039	0.0124
0.0158	44.0	9900	0.1915	0.0238	0.0072	0.0040	0.0125
0.0172	45.0	10125	0.1902	0.0246	0.0074	0.0042	0.0130
0.0145	46.0	10350	0.1842	0.0247	0.0074	0.0050	0.0122
0.0138	47.0	10575	0.1852	0.0238	0.0076	0.0047	0.0116
0.0152	48.0	10800	0.1875	0.0254	0.0075	0.0048	0.0131
0.0136	49.0	11025	0.2112	0.0244	0.0078	0.0044	0.0123
0.0137	50.0	11250	0.2049	0.0254	0.0075	0.0050	0.0129
0.0122	51.0	11475	0.2080	0.0245	0.0079	0.0036	0.0131
0.012	52.0	11700	0.2170	0.0253	0.0077	0.0046	0.0131
0.0114	53.0	11925	0.2221	0.0248	0.0074	0.0047	0.0127
0.0121	54.0	12150	0.2182	0.0240	0.0072	0.0041	0.0128
0.0123	55.0	12375	0.2091	0.0255	0.0075	0.0048	0.0131
0.0135	56.0	12600	0.2091	0.0249	0.0074	0.0051	0.0124
0.0113	57.0	12825	0.2074	0.0253	0.0076	0.0056	0.0121
0.0117	58.0	13050	0.2265	0.0250	0.0080	0.0047	0.0124
0.0115	59.0	13275	0.2364	0.0256	0.0081	0.0041	0.0134
0.0108	60.0	13500	0.2397	0.0259	0.0078	0.0047	0.0135
0.01	61.0	13725	0.2453	0.0250	0.0078	0.0043	0.0128
0.0104	62.0	13950	0.2465	0.0243	0.0075	0.0041	0.0126
0.01	63.0	14175	0.2495	0.0250	0.0079	0.0040	0.0131
0.0094	64.0	14400	0.2442	0.0255	0.0079	0.0043	0.0133
0.0098	65.0	14625	0.2485	0.0257	0.0079	0.0042	0.0136
0.0094	66.0	14850	0.2431	0.0255	0.0078	0.0045	0.0133
0.0091	67.0	15075	0.2518	0.0253	0.0078	0.0045	0.0130
0.0094	68.0	15300	0.2492	0.0250	0.0076	0.0039	0.0135
0.0095	69.0	15525	0.2444	0.0257	0.0077	0.0040	0.0139
0.0107	70.0	15750	0.2525	0.0248	0.0078	0.0038	0.0132
0.0094	71.0	15975	0.2415	0.0254	0.0077	0.0043	0.0133
0.0087	72.0	16200	0.2436	0.0252	0.0078	0.0042	0.0131
0.0086	73.0	16425	0.2546	0.0250	0.0077	0.0043	0.0130
0.0085	74.0	16650	0.2508	0.0253	0.0077	0.0043	0.0133
0.0081	75.0	16875	0.2587	0.0248	0.0078	0.0039	0.0131
0.0081	76.0	17100	0.2556	0.0250	0.0077	0.0043	0.0129
0.008	77.0	17325	0.2600	0.0250	0.0077	0.0044	0.0129
0.0081	78.0	17550	0.2650	0.0248	0.0078	0.0041	0.0129
0.0079	79.0	17775	0.2585	0.0247	0.0078	0.0044	0.0126
0.0075	80.0	18000	0.2620	0.0251	0.0077	0.0045	0.0129
0.008	81.0	18225	0.2721	0.0251	0.0078	0.0043	0.0130
0.0077	82.0	18450	0.2672	0.0246	0.0077	0.0042	0.0127
0.007	83.0	18675	0.2647	0.0250	0.0078	0.0043	0.0129
0.0078	84.0	18900	0.2787	0.0248	0.0078	0.0042	0.0128
0.0078	85.0	19125	0.2775	0.0250	0.0078	0.0041	0.0131
0.0075	86.0	19350	0.2768	0.0246	0.0077	0.0042	0.0127
0.0072	87.0	19575	0.2798	0.0247	0.0078	0.0042	0.0127
0.0069	88.0	19800	0.2764	0.0247	0.0078	0.0043	0.0126
0.0069	89.0	20025	0.2770	0.0248	0.0078	0.0043	0.0127
0.007	90.0	20250	0.2801	0.0248	0.0078	0.0042	0.0128
0.0073	91.0	20475	0.2800	0.0247	0.0078	0.0042	0.0127
0.0071	92.0	20700	0.2814	0.0247	0.0077	0.0043	0.0127
0.0077	93.0	20925	0.2799	0.0247	0.0078	0.0043	0.0126
0.0071	94.0	21150	0.2808	0.0247	0.0078	0.0043	0.0126
0.0073	95.0	21375	0.2798	0.0247	0.0078	0.0043	0.0126
0.0074	96.0	21600	0.2797	0.0247	0.0078	0.0043	0.0126
0.0071	97.0	21825	0.2798	0.0247	0.0078	0.0043	0.0126
0.0067	98.0	22050	0.2799	0.0247	0.0078	0.0043	0.0126
0.0073	99.0	22275	0.2799	0.0247	0.0078	0.0043	0.0126
0.0069	100.0	22500	0.2799	0.0247	0.0078	0.0043	0.0126

Framework versions

Transformers 4.47.1
Pytorch 2.3.0+cu118
Datasets 3.2.0
Tokenizers 0.21.0

vaibhavchavan/speaker-segmentation-fine-tuned-Redio-Audio-hi

speaker-segmentation-fine-tuned-Redio-Audio-hi

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

vaibhavchavan/
speaker-segmentation-fine-tuned-Redio-Audio-hi