speaker-segmentation-emotion

This model is a fine-tuned version of pyannote/segmentation-3.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4349
• Model Preparation Time: 0.0045
• Der: 0.2041
• False Alarm: 0.0709
• Missed Detection: 0.1190
• Confusion: 0.0141

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: 128
• eval_batch_size: 128
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 512
• total_eval_batch_size: 512
• 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_with_min_lr
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30.0

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Der	False Alarm	Missed Detection	Confusion
0.5336	1.0	2616	0.5341	0.0045	0.2408	0.0777	0.1372	0.0259
0.5071	2.0	5232	0.5025	0.0045	0.2301	0.0716	0.1374	0.0211
0.497	3.0	7848	0.4931	0.0045	0.2269	0.0679	0.1384	0.0206
0.4837	4.0	10464	0.4856	0.0045	0.2247	0.0637	0.1412	0.0197
0.471	5.0	13080	0.4726	0.0045	0.2198	0.0648	0.1373	0.0177
0.4622	6.0	15696	0.4641	0.0045	0.2172	0.0668	0.1347	0.0157
0.459	7.0	18312	0.4601	0.0045	0.2147	0.0653	0.1331	0.0163
0.4574	8.0	20928	0.4571	0.0045	0.2145	0.0645	0.1338	0.0162
0.4672	9.0	23544	0.4613	0.0045	0.2147	0.0687	0.1290	0.0170
0.4564	10.0	26160	0.4630	0.0045	0.2165	0.0684	0.1306	0.0174
0.4554	11.0	28776	0.4574	0.0045	0.2140	0.0676	0.1292	0.0172
0.4602	12.0	31392	0.4638	0.0045	0.2159	0.0698	0.1276	0.0185
0.4489	13.0	34008	0.4539	0.0045	0.2124	0.0666	0.1293	0.0165
0.4507	14.0	36624	0.4554	0.0045	0.2132	0.0695	0.1267	0.0170
0.4503	15.0	39240	0.4536	0.0045	0.2119	0.0765	0.1182	0.0172
0.4471	16.0	41856	0.4472	0.0045	0.2099	0.0698	0.1242	0.0159
0.444	17.0	44472	0.4484	0.0045	0.2095	0.0682	0.1256	0.0157
0.4423	18.0	47088	0.4413	0.0045	0.2074	0.0667	0.1262	0.0145
0.4327	19.0	49704	0.4395	0.0045	0.2061	0.0690	0.1229	0.0142
0.4357	20.0	52320	0.4375	0.0045	0.2056	0.0689	0.1226	0.0141
0.4247	21.0	54936	0.4344	0.0045	0.2039	0.0702	0.1202	0.0134
0.4334	22.0	57552	0.4378	0.0045	0.2055	0.0700	0.1208	0.0147
0.4369	23.0	60168	0.4399	0.0045	0.2061	0.0702	0.1203	0.0157
0.4341	24.0	62784	0.4376	0.0045	0.2054	0.0709	0.1197	0.0148
0.428	25.0	65400	0.4366	0.0045	0.2046	0.0693	0.1211	0.0141
0.4324	26.0	68016	0.4360	0.0045	0.2047	0.0688	0.1216	0.0142
0.4319	27.0	70632	0.4358	0.0045	0.2045	0.0718	0.1183	0.0144
0.4317	28.0	73248	0.4351	0.0045	0.2042	0.0701	0.1200	0.0141
0.4303	29.0	75864	0.4348	0.0045	0.2041	0.0709	0.1191	0.0141
0.428	30.0	78480	0.4349	0.0045	0.2041	0.0709	0.1190	0.0141

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3