wav2vec2-base-finetuned-ie

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: 1.0843
• Accuracy: 0.6511

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1746	1.0	102	1.1532	0.4811
0.9955	2.0	204	1.0070	0.5761
0.8048	3.0	306	0.9764	0.6052
0.5933	4.0	408	1.1020	0.5975
0.4987	5.0	510	1.1877	0.5984
0.401	6.0	612	1.1034	0.6334
0.3613	7.0	714	1.0424	0.6780
0.2304	8.0	816	1.1990	0.6673
0.2553	9.0	918	1.2547	0.6605
0.1737	10.0	1020	1.5981	0.6149
0.1438	11.0	1122	1.4178	0.6770
0.0917	12.0	1224	1.6137	0.6576
0.0446	13.0	1326	1.7111	0.6557
0.0397	14.0	1428	1.7771	0.6392
0.0169	15.0	1530	1.7484	0.6508

Framework versions

• Transformers 4.34.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.14.0