TrimLesson6

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: 0.2638
• Accuracy: 0.9020
• F1-score: 0.8990
• Recall-score: 0.9020
• Precision-score: 0.9085

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
3.0068	1.0	427	3.0157	0.2007	0.1021	0.2007	0.1514
2.0798	2.0	854	1.8506	0.5478	0.4726	0.5478	0.4815
1.5096	3.0	1281	1.1647	0.6963	0.6441	0.6963	0.6785
0.8799	4.0	1708	0.8187	0.7240	0.6793	0.7240	0.7035
1.8621	5.0	2135	0.7214	0.7409	0.7049	0.7409	0.7304
0.9176	6.0	2562	0.6481	0.7564	0.7249	0.7564	0.7461
0.7314	7.0	2989	0.5848	0.7695	0.7321	0.7695	0.7379
0.2837	8.0	3416	0.5256	0.7858	0.7592	0.7858	0.7798
0.5412	9.0	3843	0.5331	0.7852	0.7561	0.7852	0.7785
0.545	10.0	4270	0.5223	0.7893	0.7590	0.7893	0.7974
0.6444	11.0	4697	0.4780	0.8057	0.7896	0.8057	0.7977
0.6496	12.0	5124	0.4717	0.8049	0.7771	0.8049	0.8083
0.1724	13.0	5551	0.4521	0.8188	0.7994	0.8188	0.8357
0.4841	14.0	5978	0.4289	0.8226	0.8109	0.8226	0.8309
0.3883	15.0	6405	0.4123	0.8268	0.8075	0.8268	0.8255
0.6509	16.0	6832	0.3927	0.8467	0.8400	0.8467	0.8559
0.6592	17.0	7259	0.3711	0.8503	0.8415	0.8503	0.8617
0.2939	18.0	7686	0.3645	0.8525	0.8368	0.8525	0.8687
0.0568	19.0	8113	0.3307	0.8727	0.8675	0.8727	0.8806
0.2942	20.0	8540	0.3354	0.8715	0.8668	0.8715	0.8800
0.4429	21.0	8967	0.3063	0.8821	0.8775	0.8821	0.8892
0.1764	22.0	9394	0.2903	0.8904	0.8849	0.8904	0.9002
0.0734	23.0	9821	0.2816	0.8927	0.8873	0.8927	0.9007
0.5793	24.0	10248	0.2635	0.9077	0.9062	0.9077	0.9092
0.2896	25.0	10675	0.2638	0.9020	0.8990	0.9020	0.9085

Framework versions

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