Artanis1551/whisper_romanian

Whisper Romanian

This model is a fine-tuned version of openai/whisper-small on the Romanian voice 1.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2557
• Wer: 23.8763

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0156	0.45	250	0.2336	24.4175
0.011	0.9	500	0.2361	25.0627
0.0061	1.35	750	0.2479	27.5539
0.0084	1.8	1000	0.2491	27.3174
0.0054	2.25	1250	0.2532	21.1234
0.0041	2.7	1500	0.2557	23.8763

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0