trim-lesson7-classification

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

• Loss: 0.1884
• Accuracy: 0.9670
• F1-score: 0.9671
• Recall-score: 0.9670
• Precision-score: 0.9679

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1-score	Recall-score	Precision-score
3.8363	1.0	223	3.7557	0.1494	0.0778	0.1494	0.0756
2.7979	2.0	446	2.5827	0.6355	0.5759	0.6355	0.6166
1.6429	3.0	669	1.5768	0.9235	0.9229	0.9235	0.9280
1.659	4.0	892	0.8894	0.9440	0.9439	0.9440	0.9461
0.5305	5.0	1115	0.5156	0.9557	0.9557	0.9557	0.9573
0.3831	6.0	1338	0.3869	0.9537	0.9539	0.9537	0.9560
0.1822	7.0	1561	0.3246	0.9576	0.9576	0.9576	0.9589
0.1308	8.0	1784	0.2840	0.9523	0.9524	0.9523	0.9540
0.1087	9.0	2007	0.3299	0.9401	0.9393	0.9401	0.9460
0.0968	10.0	2230	0.2539	0.9548	0.9549	0.9548	0.9571
0.1088	11.0	2453	0.2290	0.9606	0.9606	0.9606	0.9617
0.7219	12.0	2676	0.2346	0.9606	0.9607	0.9606	0.9616
0.2103	13.0	2899	0.2119	0.9629	0.9629	0.9629	0.9640
0.0414	14.0	3122	0.2431	0.9590	0.9590	0.9590	0.9603
0.9212	15.0	3345	0.2141	0.9651	0.9651	0.9651	0.9664
0.0244	16.0	3568	0.2185	0.9620	0.9620	0.9620	0.9633
0.0468	17.0	3791	0.1949	0.9645	0.9645	0.9645	0.9655
1.2045	18.0	4014	0.1985	0.9637	0.9637	0.9637	0.9648
0.1907	19.0	4237	0.1894	0.9634	0.9635	0.9634	0.9646
0.0185	20.0	4460	0.1956	0.9640	0.9640	0.9640	0.9648
0.0159	21.0	4683	0.2118	0.9601	0.9601	0.9601	0.9610
0.0633	22.0	4906	0.1953	0.9634	0.9635	0.9634	0.9646
0.0244	23.0	5129	0.1915	0.9665	0.9665	0.9665	0.9673
0.008	24.0	5352	0.1842	0.9690	0.9690	0.9690	0.9698
0.2523	25.0	5575	0.1884	0.9670	0.9671	0.9670	0.9679

Framework versions

• Transformers 4.45.1
• Pytorch 2.3.0+cu118
• Datasets 3.0.1
• Tokenizers 0.20.0