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_15epochs_shuffled

This model is a fine-tuned version of openai/whisper-medium on an unknown 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: 5e-06
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 128
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: reduce_lr_on_plateau
lr_scheduler_warmup_steps: 500
num_epochs: 15
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Cer
0.3769	1.0	79	0.3332	10.4000
0.2757	2.0	158	0.3078	9.8627
0.2516	3.0	237	0.3062	9.8224
0.2249	4.0	316	0.3069	9.5589
0.2151	5.0	395	0.3111	9.6482
0.1959	6.0	474	0.3151	9.6568

Safetensors

Model size

0.8B params

Tensor type

F32

Base model

Finetuned

(724)

this model