steja/whisper-small-khmer

Whisper_small_Khmer

This model is a fine-tuned version of openai/whisper-small on the google/fleurs km_kh dataset. It achieves the following results on the evaluation set:

• Loss: 1.9221
• Wer: 84.9417

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: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5571	40.0	400	1.2022	89.9564
0.0088	80.0	800	1.7980	86.6669
0.0023	120.0	1200	1.9221	84.9417
0.0002	160.0	1600	2.0559	85.4326
0.0002	200.0	2000	2.0787	85.6536

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2