uaspeech-large-finetune-shorter-evals

This model is a fine-tuned version of openai/whisper-large-v3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2762

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss
0.316	0.0828	200	0.3907
0.2478	0.1242	300	0.3199
0.2129	0.1656	400	0.3282
0.1667	0.2070	500	0.3194
0.1534	0.2483	600	0.3327
0.1208	0.2897	700	0.2923
0.0987	0.3311	800	0.3048
0.103	0.3725	900	0.2841
0.0893	0.4139	1000	0.2759
0.0757	0.4553	1100	0.2625
0.068	0.4967	1200	0.2784
0.0608	0.5381	1300	0.2813
0.0404	0.5795	1400	0.2739
0.0422	0.6209	1500	0.2762

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3