whisper-tiny-khmer-aug-v6

This model is a fine-tuned version of openai/whisper-tiny on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2585
• Wer: 67.4882

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: 16
• 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: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7681	0.9994	837	0.3735	85.3089
0.32	2.0	1675	0.2922	79.2119
0.2456	2.9994	2512	0.2541	77.7850
0.2055	4.0	3350	0.2429	76.9418
0.1788	4.9994	4187	0.2487	73.0177
0.1589	6.0	5025	0.2357	68.6882
0.1419	6.9994	5862	0.2396	70.0503
0.1283	8.0	6700	0.2448	66.4829
0.1154	8.9994	7537	0.2484	68.8503
0.1046	9.9940	8370	0.2585	67.4882

Framework versions

• Transformers 4.44.0
• Pytorch 2.3.1
• Datasets 2.21.0
• Tokenizers 0.19.1