wav2vec2transformerEMR

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: 0.6501
• Accuracy: 0.7937
• Precision: 0.7945
• Recall: 0.7937
• F1: 0.7924

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.6305	0.8210	500	1.5561	0.4443	0.4495	0.4443	0.3962
1.1604	1.6420	1000	1.1252	0.6510	0.6854	0.6510	0.6491
0.9048	2.4631	1500	0.9422	0.7008	0.7202	0.7008	0.6987
0.7442	3.2841	2000	0.8200	0.7398	0.7561	0.7398	0.7358
0.6853	4.1051	2500	0.7475	0.7587	0.7646	0.7587	0.7555
0.6067	4.9261	3000	0.7000	0.7731	0.7860	0.7731	0.7748
0.5184	5.7471	3500	0.6890	0.7801	0.7853	0.7801	0.7778
0.4781	6.5681	4000	0.6983	0.7768	0.7888	0.7768	0.7752
0.4078	7.3892	4500	0.6654	0.7916	0.7979	0.7916	0.7913
0.4012	8.2102	5000	0.6759	0.7908	0.8003	0.7908	0.7897
0.3964	9.0312	5500	0.6501	0.7937	0.7945	0.7937	0.7924
0.315	9.8522	6000	0.6744	0.7887	0.7932	0.7887	0.7866

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Tokenizers 0.20.3