ihanif/whisper-medium-pashto-3e-7

openai/whisper-medium

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

• Loss: 1.0063
• Wer: 63.1053

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: 3e-07
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10
• training_steps: 700
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.4728	14.29	100	1.4429	83.5942
1.0344	28.57	200	1.1584	69.9706
0.9318	42.86	300	1.1061	67.8394
0.8882	57.14	400	1.0769	66.4290
0.8609	71.43	500	1.0575	66.1965
0.8262	85.71	600	1.0214	63.6274
0.7989	100.0	700	1.0063	63.1053

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2