xuliu15/Frisian_32r_LoRA_10h

LoRA-Frisian-10h

This model is a fine-tuned version of openai/whisper-small on the mozilla-foundation/common_voice_6_1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4038

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: 0.001
• 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: 50
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5243	1.0	935	0.6300
0.432	2.0	1870	0.5607
0.3525	3.0	2805	0.5192
0.2719	4.0	3740	0.4961
0.2251	5.0	4675	0.4779
0.1761	6.0	5610	0.4505
0.1344	7.0	6545	0.4435
0.1019	8.0	7480	0.4414
0.0717	9.0	8415	0.4197
0.0434	10.0	9350	0.4149
0.023	11.0	10285	0.4077
0.0161	12.0	11220	0.4067
0.0075	13.0	12155	0.3982
0.0066	14.0	13090	0.4005
0.0034	15.0	14025	0.4038

Framework versions

• PEFT 0.10.0
• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1