wav2vec2-base_toy_train_data_random_noise

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0909
• Wer: 0.7351

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.128	2.1	250	3.5052	1.0
3.0423	4.2	500	2.9312	1.0
1.4109	6.3	750	1.2618	0.8915
0.9132	8.4	1000	1.1074	0.8436
0.7146	10.5	1250	1.0397	0.7876
0.5418	12.6	1500	1.0359	0.7662
0.4649	14.7	1750	1.0469	0.7467
0.4127	16.8	2000	1.0655	0.7404
0.3881	18.9	2250	1.0909	0.7351

Framework versions

• Transformers 4.17.0
• Pytorch 1.11.0+cu102
• Datasets 2.0.0
• Tokenizers 0.11.6