thisiskeithkwan/cantomed-base

Whisper medium 12

This model is a fine-tuned version of openai/whisper-medium on the mozilla-foundation/common_voice_11_0 dataset. It achieves the following results on the evaluation set:

• Loss: 1.3270
• Cer: 42.0122

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 8000

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.8931	1.52	1000	1.0926	48.9439
0.3041	3.03	2000	1.1069	49.5474
0.1319	4.55	3000	1.1925	45.4016
0.0324	6.06	4000	1.2592	44.3186
0.0245	7.58	5000	1.3014	44.2359
0.0061	9.09	6000	1.3185	43.3472
0.0031	10.61	7000	1.3266	42.2767
0.0007	12.12	8000	1.3270	42.0122

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3