audioclass-alpha

This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1124
• Accuracy: 0.9660

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.4324	1.0	62	3.4330	0.0295
3.4178	2.0	124	3.4067	0.0839
3.3067	3.0	186	3.2018	0.3107
3.0013	4.0	248	2.7640	0.5261
2.5488	5.0	310	2.3188	0.6644
2.1833	6.0	372	1.9013	0.7687
1.7949	7.0	434	1.5320	0.8141
1.5859	8.0	496	1.2519	0.8413
1.2774	9.0	558	1.0155	0.8662
1.1146	10.0	620	0.8348	0.8776
0.9276	11.0	682	0.7070	0.8844
0.7634	12.0	744	0.5845	0.8889
0.726	13.0	806	0.5491	0.8866
0.6325	14.0	868	0.4927	0.8707
0.5525	15.0	930	0.4065	0.8866
0.5051	16.0	992	0.4063	0.8798
0.4543	17.0	1054	0.4166	0.8685
0.4138	18.0	1116	0.3328	0.8889
0.4133	19.0	1178	0.2988	0.8934
0.4087	20.0	1240	0.3092	0.8934
0.3402	21.0	1302	0.2600	0.9002
0.3052	22.0	1364	0.2779	0.8957
0.2792	23.0	1426	0.2318	0.9274
0.3357	24.0	1488	0.2348	0.9274
0.2602	25.0	1550	0.2928	0.9274
0.2582	26.0	1612	0.2410	0.9388
0.1906	27.0	1674	0.2294	0.9433
0.1937	28.0	1736	0.2176	0.9456
0.3112	29.0	1798	0.1707	0.9501
0.1854	30.0	1860	0.1798	0.9501
0.2662	31.0	1922	0.1650	0.9546
0.1892	32.0	1984	0.1636	0.9524
0.1652	33.0	2046	0.1688	0.9524
0.2581	34.0	2108	0.1324	0.9615
0.2007	35.0	2170	0.1400	0.9592
0.1368	36.0	2232	0.1510	0.9569
0.1397	37.0	2294	0.1168	0.9637
0.1604	38.0	2356	0.1203	0.9615
0.1638	39.0	2418	0.1224	0.9637
0.1892	40.0	2480	0.1148	0.9592
0.1647	41.0	2542	0.1004	0.9637
0.1337	42.0	2604	0.1124	0.9660
0.102	43.0	2666	0.1021	0.9637
0.1293	44.0	2728	0.1053	0.9615
0.2035	45.0	2790	0.1033	0.9637
0.1222	46.0	2852	0.1045	0.9615
0.1393	47.0	2914	0.1043	0.9615
0.1271	48.0	2976	0.1055	0.9615
0.1618	49.0	3038	0.1057	0.9615
0.1536	50.0	3100	0.1046	0.9615

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1