Pageee/DistilFT-Frisian-10hd

DistilFT-Frisian-10h

This model is a fine-tuned version of distil-small.en on the mozilla-foundation/common_voice_6_fy_NL dataset. It achieves the following results on the evaluation set:

• Loss: 0.5755
• Wer: 26.9114

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: 0.0001
• train_batch_size: 8
• 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: 300
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.9504	0.5348	500	1.0939	57.4122
0.4656	1.0695	1000	0.8241	45.7316
0.4533	1.6043	1500	0.7285	41.3474
0.1745	2.1390	2000	0.6875	37.7009
0.1701	2.6738	2500	0.6261	34.7603
0.0709	3.2086	3000	0.6566	33.4415
0.0731	3.7433	3500	0.5880	30.5650
0.0234	4.2781	4000	0.5949	28.8754
0.0192	4.8128	4500	0.5799	27.7063
0.0038	5.3476	5000	0.5755	26.9114

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1