distil_noisy_teacher

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

• Loss: 5.4992
• Wer: 0.0478

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: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 256
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
32.3439	22.22	1000	8.8146	0.1150
16.0955	44.44	2000	7.5152	0.0890
8.5751	66.67	3000	6.0369	0.0602
4.3989	88.89	4000	5.4992	0.0478

Framework versions

• Transformers 4.25.1
• Pytorch 1.12.1
• Datasets 2.7.1
• Tokenizers 0.11.0