StefanJevtic63/whisper-large-v2-sr-lora-learning-rate-0.0005

Whisper - Serbian Model

This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1707

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: 0.0005
• train_batch_size: 4
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 400
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2354	0.0088	500	0.2467
0.195	0.0176	1000	0.2312
0.1977	0.0264	1500	0.2173
0.178	0.0352	2000	0.2052
0.1676	0.0440	2500	0.1929
0.2103	0.0529	3000	0.1852
0.1675	0.0617	3500	0.1755
0.1888	0.0705	4000	0.1707

Framework versions

• PEFT 0.13.2
• Transformers 4.46.3
• Pytorch 2.5.1
• Datasets 3.0.0
• Tokenizers 0.20.3