Pak-Speech-Processing/urdu-emotions-wav2vec2-large-xlsr

urdu-emotions

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0077
• Accuracy: 1.0

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	8	1.0905	0.4833
1.0739	2.0	16	1.0736	0.5
1.0341	3.0	24	0.9888	0.75
0.9797	4.0	32	0.9144	0.7667
0.8716	5.0	40	0.8531	0.75
0.8716	6.0	48	0.7572	0.75
0.7905	7.0	56	0.6556	0.7333
0.6837	8.0	64	0.5976	0.7167
0.6611	9.0	72	0.5298	0.7667
0.5744	10.0	80	0.5014	0.8
0.5744	11.0	88	0.4616	0.8667
0.5402	12.0	96	0.3964	0.9667
0.4185	13.0	104	0.3536	0.8833
0.4138	14.0	112	0.2818	0.9167
0.3409	15.0	120	0.2554	0.9333
0.3409	16.0	128	0.1434	0.9833
0.2923	17.0	136	0.1658	0.95
0.2412	18.0	144	0.0775	1.0
0.2038	19.0	152	0.0766	1.0
0.1618	20.0	160	0.0680	1.0
0.1618	21.0	168	0.0517	1.0
0.2041	22.0	176	0.0344	1.0
0.1238	23.0	184	0.0313	1.0
0.1179	24.0	192	0.0269	1.0
0.1738	25.0	200	0.0248	1.0
0.1738	26.0	208	0.0213	1.0
0.1265	27.0	216	0.0363	0.9833
0.1503	28.0	224	0.1058	0.9667
0.1201	29.0	232	0.0270	1.0
0.0484	30.0	240	0.1284	0.9667
0.0484	31.0	248	0.1702	0.95
0.2394	32.0	256	0.0144	1.0
0.1625	33.0	264	0.0145	1.0
0.0999	34.0	272	0.0172	1.0
0.0406	35.0	280	0.0106	1.0
0.0406	36.0	288	0.0098	1.0
0.0374	37.0	296	0.0097	1.0
0.0268	38.0	304	0.0100	1.0
0.0279	39.0	312	0.0111	1.0
0.031	40.0	320	0.0099	1.0
0.031	41.0	328	0.0096	1.0
0.0412	42.0	336	0.0096	1.0
0.0218	43.0	344	0.0087	1.0
0.026	44.0	352	0.0079	1.0
0.0122	45.0	360	0.0081	1.0
0.0122	46.0	368	0.0087	1.0
0.0328	47.0	376	0.0088	1.0
0.0264	48.0	384	0.0084	1.0
0.0339	49.0	392	0.0078	1.0
0.0257	50.0	400	0.0077	1.0

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0
• Datasets 2.1.0
• Tokenizers 0.12.1