arbml/whisper-small-ar

openai/whisper-small

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: 0.9750
• Wer: 21.3693

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: 64
• 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: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3559	0.1	1000	0.9147	29.3252
0.3154	0.2	2000	1.1353	26.5718
0.359	0.3	3000	0.9208	25.3987
0.273	0.4	4000	0.9591	24.3877
0.2326	0.5	5000	0.9207	21.9052
0.2992	1.04	6000	0.9445	22.4556
0.2265	1.14	7000	0.9660	21.2230
0.2059	1.24	8000	0.9785	20.9551
0.2239	1.34	9000	0.9637	21.6300
0.2163	1.44	10000	0.9750	21.3693

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2