arjunshajitech/whisper-small-malayalam

Whisper Small Malayalam - Arjun Shaji

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6067
• Wer: 85.2874

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.1903	3.7037	100	1.1262	100.0
0.473	7.4074	200	0.5343	100.9195
0.1263	11.1111	300	0.4247	91.7241
0.0335	14.8148	400	0.5135	91.7241
0.0262	18.5185	500	0.5317	91.7241
0.0135	22.2222	600	0.5361	86.2069
0.0067	25.9259	700	0.5448	84.5977
0.0016	29.6296	800	0.6192	88.0460
0.0003	33.3333	900	0.5992	84.8276
0.0002	37.0370	1000	0.6067	85.2874

Framework versions

• Transformers 4.41.0
• Pytorch 2.1.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1