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-kcc-dec

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
2.5602	1.0	973	1.0775	135.5467
1.4622	2.0	1946	0.7954	106.9363
1.2231	3.0	2919	0.7036	91.7134
1.0914	4.0	3892	0.6653	80.9836
0.993	5.0	4865	0.6408	72.1553
0.9151	6.0	5838	0.6325	64.6150
0.8503	7.0	6811	0.6252	61.3300
0.7907	8.0	7784	0.6202	64.5663
0.7387	9.0	8757	0.6344	57.2715
0.6895	9.9902	9720	0.6399	58.9179

Safetensors

Model size

37.8M params

Tensor type

F32

Base model

Finetuned

this model