TrimLesson7

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: 0.5436
• Accuracy: 0.8708
• F1-score: 0.8706
• Recall-score: 0.8708
• Precision-score: 0.8752

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: 20
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1-score	Recall-score	Precision-score
2.8852	1.0	711	2.8513	0.3919	0.3173	0.3919	0.3569
1.7162	2.0	1422	1.7808	0.6128	0.5693	0.6128	0.6253
1.5847	3.0	2133	1.1873	0.7367	0.7167	0.7367	0.7467
0.6171	4.0	2844	0.8186	0.8165	0.8141	0.8165	0.8366
0.5935	5.0	3555	0.6720	0.8390	0.8405	0.8390	0.8540
0.7741	6.0	4266	0.6298	0.8279	0.8268	0.8279	0.8441
0.3842	7.0	4977	0.5501	0.8423	0.8409	0.8423	0.8521
1.1043	8.0	5688	0.5400	0.8409	0.8403	0.8409	0.8552
0.8919	9.0	6399	0.5461	0.8465	0.8447	0.8465	0.8578
0.3236	10.0	7110	0.5841	0.8404	0.8409	0.8404	0.8530
1.1149	11.0	7821	0.5966	0.8345	0.8317	0.8345	0.8464
0.6224	12.0	8532	0.5362	0.8623	0.8635	0.8623	0.8736
0.5349	13.0	9243	0.5442	0.8570	0.8575	0.8570	0.8685
0.4529	14.0	9954	0.4945	0.8664	0.8671	0.8664	0.8752
0.5113	15.0	10665	0.4864	0.8678	0.8683	0.8678	0.8754
0.138	16.0	11376	0.5388	0.8623	0.8627	0.8623	0.8707
0.363	17.0	12087	0.5189	0.8689	0.8692	0.8689	0.8750
0.6955	18.0	12798	0.5180	0.8686	0.8683	0.8686	0.8739
0.0167	19.0	13509	0.5636	0.8650	0.8649	0.8650	0.8711
0.1746	20.0	14220	0.5436	0.8708	0.8706	0.8708	0.8752

Framework versions

• Transformers 4.45.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.20.0