You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

whisper-tiny-khmer-v9-aug

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

More information needed

More information needed

More information needed

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant
lr_scheduler_warmup_steps: 1000
num_epochs: 10

Training Loss	Epoch	Step	Validation Loss	Wer
1.0755	1.0	951	0.6682	103.7374
0.5373	2.0	1902	0.3144	93.8680
0.373	3.0	2853	0.2535	93.7177
0.3087	4.0	3804	0.2281	84.5244
0.272	5.0	4755	0.2164	87.6233
0.2442	6.0	5706	0.2020	80.8996
0.2236	7.0	6657	0.1988	77.9885
0.2068	8.0	7608	0.1919	76.5706
0.1922	9.0	8559	0.1934	77.1903
0.1795	10.0	9510	0.1920	76.5330

Safetensors

Model size

37.8M params

Tensor type

F32

Base model

Finetuned

this model