liqi03/whisper-large-v3-tr-fleurs

Whisper Large V3 tr Fleurs - Chee Li

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

• Loss: 0.1432
• Wer: 649.9222

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: 16
• 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: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0466	2.7933	500	0.1060	147.9932
0.006	5.5866	1000	0.1208	481.1605
0.0017	8.3799	1500	0.1291	602.0769
0.0012	11.1732	2000	0.1288	627.3647
0.0002	13.9665	2500	0.1382	641.4203
0.0001	16.7598	3000	0.1411	647.7520
0.0001	19.5531	3500	0.1426	642.9294
0.0001	22.3464	4000	0.1432	649.9222

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1