base

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: 0.6689
• Wer: 31.2258
• Cer: 13.9975

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: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 2700

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
4.3822	1.4	500	2.9318	100.0	99.3341
2.7693	2.81	1000	1.7387	90.0629	50.4533
0.9915	4.21	1500	0.7714	41.5179	18.1139
0.6189	5.62	2000	0.7052	35.2091	15.8310
0.4696	7.02	2500	0.6689	31.2258	13.9975

Framework versions

• Transformers 4.39.3
• Pytorch 1.12.1+cu116
• Datasets 2.18.0
• Tokenizers 0.15.2