CngFSt3T_small_model

This model is a fine-tuned version of openai/whisper-small on the Marcusxx/CngFSt3T dataset. It achieves the following results on the evaluation set:

• Loss: 0.1722
• Cer: 158.1261

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-06
• 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: 250
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.4693	2.2321	1000	0.5347	45.6669
0.2295	4.4643	2000	0.3416	101.5993
0.1157	6.6964	3000	0.2390	136.4398
0.071	8.9286	4000	0.1865	134.9996
0.0579	11.1607	5000	0.1722	158.1261

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.2+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1