whisper-small-full-data-language-v2-20ep

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1536

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: 1.25e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: tpu
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5000
• training_steps: 63840

Training results

Training Loss	Epoch	Step	Validation Loss
0.1476	1.57	5000	0.2646
0.0995	3.13	10000	0.2161
0.0835	4.7	15000	0.1959
0.0706	6.27	20000	0.1845
0.063	7.83	25000	0.1750
0.0523	9.4	30000	0.1691
0.0497	10.97	35000	0.1629
0.0407	12.53	40000	0.1606
0.0345	14.1	45000	0.1583
0.0307	15.67	50000	0.1569
0.0266	17.23	55000	0.1555
0.0244	18.8	60000	0.1540

Framework versions

• Transformers 4.28.1
• Pytorch 2.1.0a0+gitcc01568
• Datasets 2.13.1
• Tokenizers 0.13.3