multiclass-fz-enc-base

This model is a fine-tuned version of openai/whisper-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1236
• Wer: 35.4226
• Cer: 22.0113

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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.0855	4.5872	500	0.2362	45.9639	24.1240
0.0062	9.1743	1000	0.1462	44.1121	24.7595
0.0008	13.7615	1500	0.1357	36.3248	20.2508
0.0005	18.3486	2000	0.1312	35.3751	19.7613
0.0003	22.9358	2500	0.1286	35.0902	19.4091
0.0002	27.5229	3000	0.1266	34.9003	20.1477
0.0002	32.1101	3500	0.1254	35.8974	22.1402
0.0002	36.6972	4000	0.1246	36.2298	23.1106
0.0001	41.2844	4500	0.1238	35.4701	22.7843
0.0001	45.8716	5000	0.1236	35.4226	22.0113

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0