wav2vec2-base-finetuned-iemocap-fin

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: 1.1760
• Accuracy: 0.5839
• F1: 0.5773

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	F1
1.2283	1.0	102	1.2181	0.4840	0.4756
1.124	2.0	204	1.1143	0.5015	0.4808
1.062	3.0	306	1.1103	0.5189	0.5067
0.9863	4.0	408	1.0813	0.5189	0.5152
0.9689	5.0	510	1.0689	0.5403	0.5318
0.8722	6.0	612	1.0976	0.5296	0.4992
0.8757	7.0	714	1.0409	0.5606	0.5518
0.8548	8.0	816	1.0479	0.5694	0.5636
0.838	9.0	918	1.1700	0.5422	0.5109
0.7536	10.0	1020	1.0743	0.5674	0.5681
0.6557	11.0	1122	1.1487	0.5616	0.5495
0.6193	12.0	1224	1.1239	0.5849	0.5815
0.5742	13.0	1326	1.1793	0.5742	0.5617
0.5717	14.0	1428	1.1548	0.5868	0.5809
0.5929	15.0	1530	1.1760	0.5839	0.5773

Framework versions

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