wav2vec2-base-960h-EMOPIA-10sec-full-50epoc

This model is a fine-tuned version of facebook/wav2vec2-base-960h on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2688
• Accuracy: 0.8630

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• 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
1.2025	1.0	2248	1.3446	0.5196
1.415	2.0	4496	1.6350	0.6032
1.4176	3.0	6744	1.6250	0.6584
1.384	4.0	8992	1.3694	0.7242
1.3658	5.0	11240	1.4331	0.7100
1.2763	6.0	13488	1.3311	0.7438
1.2175	7.0	15736	1.2727	0.7580
1.1276	8.0	17984	1.4520	0.7331
1.1053	9.0	20232	1.2134	0.7722
1.0314	10.0	22480	1.2143	0.7829
1.0029	11.0	24728	1.3312	0.7811
0.9108	12.0	26976	1.2228	0.8025
0.8335	13.0	29224	1.1526	0.8078
0.8514	14.0	31472	0.9904	0.8203
0.7389	15.0	33720	1.3000	0.8025
0.6993	16.0	35968	1.0873	0.8203
0.6177	17.0	38216	1.0856	0.8327
0.641	18.0	40464	1.3224	0.7972
0.611	19.0	42712	1.1800	0.8292
0.5744	20.0	44960	1.2937	0.8096
0.5008	21.0	47208	1.1565	0.8416
0.4396	22.0	49456	1.3663	0.8149
0.4313	23.0	51704	1.3267	0.8221
0.3954	24.0	53952	1.1824	0.8470
0.4217	25.0	56200	1.5586	0.8043
0.3797	26.0	58448	1.1746	0.8523
0.358	27.0	60696	1.1937	0.8452
0.2963	28.0	62944	1.4036	0.8310
0.3338	29.0	65192	1.3134	0.8505
0.2565	30.0	67440	1.4806	0.8345
0.2798	31.0	69688	1.5173	0.8310
0.2674	32.0	71936	1.5758	0.8132
0.2334	33.0	74184	1.3401	0.8559
0.2352	34.0	76432	1.2717	0.8470
0.2406	35.0	78680	1.6163	0.8256
0.2208	36.0	80928	1.3815	0.8505
0.1796	37.0	83176	1.3929	0.8577
0.2127	38.0	85424	1.5271	0.8274
0.1748	39.0	87672	1.5069	0.8416
0.1612	40.0	89920	1.3966	0.8470
0.1757	41.0	92168	1.4628	0.8470
0.1664	42.0	94416	1.3363	0.8523
0.1313	43.0	96664	1.4388	0.8434
0.1272	44.0	98912	1.3670	0.8630
0.1127	45.0	101160	1.4244	0.8541
0.1062	46.0	103408	1.3812	0.8541
0.0924	47.0	105656	1.4448	0.8541
0.0998	48.0	107904	1.3051	0.8683
0.1055	49.0	110152	1.2630	0.8701
0.1073	50.0	112400	1.2688	0.8630

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu118
• Datasets 3.0.1
• Tokenizers 0.20.0