wav2vec2-base-finetuned-iemocap-fin2

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: 1.1912
• Accuracy: 0.5597
• F1: 0.5453

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
1.2603	1.0	102	1.2581	0.4617	0.4105
1.1338	2.0	204	1.1471	0.4801	0.4369
1.0899	3.0	306	1.1386	0.4782	0.4459
1.0501	4.0	408	1.0894	0.5218	0.5096
0.9892	5.0	510	1.0778	0.5422	0.5339
0.8943	6.0	612	1.1394	0.5141	0.4730
0.9112	7.0	714	1.0634	0.5529	0.5379
0.8688	8.0	816	1.0726	0.5664	0.5576
0.8807	9.0	918	1.2264	0.5209	0.4822
0.8027	10.0	1020	1.0469	0.5839	0.5843
0.7069	11.0	1122	1.1171	0.5587	0.5398
0.6508	12.0	1224	1.1889	0.5480	0.5292
0.6406	13.0	1326	1.1800	0.5664	0.5501
0.6072	14.0	1428	1.1841	0.5558	0.5413
0.6277	15.0	1530	1.1912	0.5597	0.5453

Framework versions

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