wav2vec2-base-alffaamharic-google-colab

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

• Loss: 0.6124
• Wer: 0.3502

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: 0.0005
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1622	1.44	200	0.4605	0.3455
0.1619	2.88	400	0.5327	0.3658
0.1566	4.32	600	0.5235	0.3834
0.246	5.76	800	0.5781	0.4329
0.3211	7.19	1000	0.6142	0.4535
0.2705	8.63	1200	0.6172	0.4519
0.2113	10.07	1400	0.6140	0.4111
0.2642	11.51	1600	0.6020	0.4527
0.2881	12.95	1800	0.5931	0.4393
0.1801	14.39	2000	0.6565	0.4262
0.1494	15.83	2200	0.5732	0.4003
0.1684	17.27	2400	0.6419	0.4071
0.1805	18.71	2600	0.5973	0.4005
0.1518	20.14	2800	0.5846	0.3821
0.0821	21.58	3000	0.6149	0.3764
0.1049	23.02	3200	0.5965	0.3724
0.1229	24.46	3400	0.6032	0.3671
0.0848	25.9	3600	0.6005	0.3605
0.0724	27.34	3800	0.6258	0.3634
0.1149	28.78	4000	0.6124	0.3502

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3