whisper-tamil-vulnerablee

This model is a fine-tuned version of HariprasathSB/whisper-tamil-vulnerable on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1757
• Wer: 76.6682

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: 1e-05
• train_batch_size: 8
• 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: 200
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0216	1.7544	200	1.0816	78.3139
0.0191	3.5088	400	1.0777	79.3327
0.0069	5.2632	600	1.1236	77.1048
0.003	7.0175	800	1.1772	78.3699
0.0004	8.7719	1000	1.1757	76.6682

Framework versions

• Transformers 4.41.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1