monsub-subtitle-v1

This model is a fine-tuned version of Tsedee/whisper-large-v3-turbo-mn-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0699
• Wer: 10.29

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• training_steps: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0447	0.8937	250	0.0659	9.82
0.0237	1.7873	500	0.0637	9.89
0.0153	2.6810	750	0.0706	10.39
0.0064	3.5746	1000	0.0693	9.93
0.0027	4.4683	1250	0.0667	10.0
0.0015	5.3619	1500	0.0699	10.29

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.4.0+cu124
• Datasets 2.21.0
• Tokenizers 0.19.1