deepdml/whisper-small-mix-it

Whisper Small Mixed-Italian

This model is a fine-tuned version of openai/whisper-small on the it datasets:

• mozilla-foundation/common_voice_17_0
• google/fleurs
• facebook/multilingual_librispeech

It achieves the following results on the evaluation set:

• Loss: 0.1909
• Wer: 10.5875

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-05
• train_batch_size: 64
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2213	0.2	1000	0.2407	13.4605
0.1582	0.4	2000	0.2143	12.2642
0.1913	0.6	3000	0.2022	11.2328
0.1538	0.8	4000	0.1951	11.1187
0.1286	1.0	5000	0.1909	10.5875

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1