CS337_finetune_wav2vec_base

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1925
• Accuracy: 0.6574

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.97	28	2.0354	0.1612
No log	1.98	57	1.9414	0.3769
No log	2.99	86	1.7627	0.4911
1.8976	4.0	115	1.6221	0.5799
1.8976	4.97	143	1.5168	0.6117
1.8976	5.98	172	1.4421	0.6345
1.8976	6.99	201	1.3921	0.6129
1.4588	8.0	230	1.3384	0.6332
1.4588	8.97	258	1.3194	0.6371
1.4588	9.98	287	1.2695	0.6536
1.4588	10.99	316	1.2573	0.6371
1.2555	12.0	345	1.2339	0.6574
1.2555	12.97	373	1.2325	0.6472
1.2555	13.98	402	1.2141	0.6586
1.2555	14.99	431	1.2057	0.6650
1.1606	16.0	460	1.2044	0.6459
1.1606	16.97	488	1.2027	0.6459
1.1606	17.98	517	1.1929	0.6548
1.1606	18.99	546	1.1929	0.6574
1.1606	19.48	560	1.1925	0.6574

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0