pphuc25/whisper-base-full-data-language-v1-20ep

whisper-base-full-data-v4

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

• Loss: 0.2221

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
• distributed_type: tpu
• num_devices: 8
• gradient_accumulation_steps: 2
• 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
0.2569	1.57	5000	0.4138
0.1774	3.13	10000	0.3295
0.1485	4.7	15000	0.2936
0.1316	6.27	20000	0.2745
0.1187	7.83	25000	0.2599
0.1037	9.4	30000	0.2496
0.1011	10.97	35000	0.2407
0.0905	12.53	40000	0.2356
0.0827	14.1	45000	0.2316
0.0789	15.67	50000	0.2278
0.0741	17.23	55000	0.2248
0.0714	18.8	60000	0.2228

Framework versions

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