test-model

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: 0.9852
• Accuracy: 0.9663

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9771	32	2.6002	0.3131
No log	1.9847	65	2.2128	0.3973
No log	2.9924	98	1.8853	0.4949
2.4861	4.0	131	1.5879	0.6162
2.4861	4.9771	163	1.3575	0.7037
2.4861	5.9847	196	1.2166	0.8182
1.7056	6.9924	229	1.0793	0.8923
1.7056	8.0	262	0.9852	0.9663
1.7056	8.9771	294	0.9393	0.9562
1.276	9.7710	320	0.9227	0.9495

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1