iewav2vec2-finetuned-on-shEMO_speech

This model is a fine-tuned version of minoosh/wav2vec2-base-finetuned-ie on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8165
• Accuracy: 0.82

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
1.0124	1.0	75	0.9200	0.7033
0.7081	2.0	150	0.6291	0.8167
0.5352	3.0	225	0.5825	0.83
0.3974	4.0	300	0.6690	0.8
0.2653	5.0	375	0.5241	0.8567
0.1613	6.0	450	0.5175	0.86
0.1707	7.0	525	0.6105	0.85
0.1788	8.0	600	0.6798	0.8433
0.1055	9.0	675	0.6840	0.85
0.107	10.0	750	0.6162	0.87
0.0429	11.0	825	0.6344	0.8733
0.0358	12.0	900	0.7944	0.85
0.028	13.0	975	0.7262	0.8667
0.0106	14.0	1050	0.6811	0.87
0.0118	15.0	1125	0.7056	0.8667

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1