wav2vec2-large-robust-finetuned-ie

This model is a fine-tuned version of facebook/wav2vec2-large-robust on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1133
• Accuracy: 0.5536

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
1.3699	1.0	102	1.3843	0.2502
1.2027	2.0	204	1.1851	0.4200
1.047	3.0	306	1.1356	0.4520
1.0414	4.0	408	1.1702	0.4627
0.9885	5.0	510	1.0295	0.5354
0.9873	6.0	612	1.0988	0.5228
0.9309	7.0	714	1.1347	0.5257
0.8401	8.0	816	1.1502	0.5286
0.8253	9.0	918	1.0792	0.5577
0.8741	10.0	1020	1.2591	0.5267
0.8177	11.0	1122	1.3007	0.5141
0.7633	12.0	1224	1.1962	0.5509
0.8185	13.0	1326	1.1022	0.5984
0.7481	14.0	1428	1.1741	0.5694
0.719	15.0	1530	1.1645	0.5781

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2