ar-eng-autonote-turbo-exp-1

This model is a fine-tuned version of openai/whisper-large-v3-turbo on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.5087
• Bleu: 21.2006

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: 5e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
4.0421	1.0	147	3.1522	11.1208
3.0199	2.0	294	2.7301	12.1175
2.306	3.0	441	2.5164	14.4358
2.0018	4.0	588	2.4287	16.7516
1.6744	5.0	735	2.3919	19.6511
1.5185	6.0	882	2.3922	18.5162
1.3278	7.0	1029	2.4099	21.0711
1.245	8.0	1176	2.4405	18.8528
1.1109	9.0	1323	2.4835	20.8498
0.9965	10.0	1470	2.5087	21.2006

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0