TrimLesson8-9

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: 0.0802
• Accuracy: 0.9873
• F1-score: 0.9873
• Recall-score: 0.9873
• Precision-score: 0.9874

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: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• 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
2.7148	1.0	156	2.7280	0.2544	0.1660	0.2544	0.1877
1.4541	2.0	312	1.5075	0.7997	0.7575	0.7997	0.7669
0.4792	3.0	468	0.4390	0.9650	0.9647	0.9650	0.9667
0.1559	4.0	624	0.1896	0.9666	0.9662	0.9666	0.9701
0.0931	5.0	780	0.1697	0.9626	0.9632	0.9626	0.9659
0.0333	6.0	936	0.1022	0.9793	0.9792	0.9793	0.9806
0.0114	7.0	1092	0.0693	0.9873	0.9873	0.9873	0.9875
0.0112	8.0	1248	0.0997	0.9801	0.9798	0.9801	0.9810
0.0126	9.0	1404	0.0788	0.9849	0.9847	0.9849	0.9855
0.0052	10.0	1560	0.0661	0.9881	0.9880	0.9881	0.9884
0.0055	11.0	1716	0.0665	0.9865	0.9865	0.9865	0.9866
0.0031	12.0	1872	0.0777	0.9849	0.9848	0.9849	0.9853
0.0036	13.0	2028	0.1021	0.9801	0.9801	0.9801	0.9811
0.0018	14.0	2184	0.1962	0.9666	0.9662	0.9666	0.9749
0.0032	15.0	2340	0.1191	0.9809	0.9810	0.9809	0.9814
0.003	16.0	2496	0.0956	0.9817	0.9816	0.9817	0.9819
0.0017	17.0	2652	0.0735	0.9865	0.9865	0.9865	0.9867
0.0017	18.0	2808	0.0844	0.9825	0.9825	0.9825	0.9832
0.0023	19.0	2964	0.0809	0.9881	0.9881	0.9881	0.9883
0.0019	20.0	3120	0.0932	0.9833	0.9834	0.9833	0.9836
0.0011	21.0	3276	0.0942	0.9849	0.9849	0.9849	0.9852
0.0015	22.0	3432	0.0836	0.9857	0.9857	0.9857	0.9859
0.0013	23.0	3588	0.0929	0.9865	0.9865	0.9865	0.9867
0.0009	24.0	3744	0.0798	0.9873	0.9873	0.9873	0.9874
0.0016	25.0	3900	0.0802	0.9873	0.9873	0.9873	0.9874

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu118
• Datasets 2.20.0
• Tokenizers 0.20.0