Makkoen/whisper-medium.en-cit-do015-wd0-lr1e-06-SF-300

./300

This model is a fine-tuned version of openai/whisper-medium.en on the 300 SF 1000 dataset. It achieves the following results on the evaluation set:

• Loss: 0.75
• Wer Ortho: 33.0175
• Wer: 21.3491

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: 1e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• 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: 200
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
1.6187	5.2632	100	1.1230	41.5087	30.1041
0.7964	10.5263	200	0.8076	31.8878	20.2368
0.5424	15.7895	300	0.7627	31.7055	20.0933
0.4204	21.0526	400	0.7495	33.1268	21.2056
0.3666	26.3158	500	0.75	33.0175	21.3491

Framework versions

• Transformers 4.44.0
• Pytorch 1.13.1+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1