whisper-tiny-v2-ta_tf32_adamBetas

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

• Loss: 0.3434
• Wer: 42.5666

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: 1.5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5668	0.21	300	0.5259	57.9177
0.4153	0.42	600	0.4277	74.0920
0.3589	0.63	900	0.3960	51.7676
0.3132	0.84	1200	0.3748	49.5400
0.2872	1.05	1500	0.3607	42.9540
0.2793	1.26	1800	0.3537	44.9395
0.2702	1.47	2100	0.3488	43.2446
0.2708	1.68	2400	0.3434	42.5666

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 1.12.1
• Datasets 2.11.0
• Tokenizers 0.13.3