test_wav2vec2_ESC50

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: 1.0326
• Accuracy: 0.7104

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: 3e-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
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1292	0.9863	54	2.1314	0.2576
1.7714	1.9909	109	1.7812	0.4087
1.5023	2.9954	164	1.6238	0.4951
1.3402	4.0	219	1.3518	0.6022
1.1629	4.9863	273	1.2371	0.6440
1.0466	5.9909	328	1.1422	0.6789
0.9478	6.9954	383	1.0848	0.6875
0.879	8.0	438	1.1490	0.6577
0.8719	8.9863	492	1.0175	0.7121
0.8069	9.8630	540	1.0326	0.7104

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1