asr-amharic-phoneme-based-39-3

This model is a fine-tuned version of Samuael/asr-amharic-phoneme-based-39-3 on the None dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.4962
• eval_wer: 0.4152
• eval_phoneme_cer: 0.0897
• eval_cer: 0.1396
• eval_runtime: 22.8954
• eval_samples_per_second: 15.68
• eval_steps_per_second: 1.965
• epoch: 2.94
• step: 1000

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: 5e-05
• 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: 100
• num_epochs: 30

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0