libri-alpha-0.75-Temp-1-attention-3-layers-distil-with-6-layers-att-take-4

This model is a fine-tuned version of rohitp1/libri-alpha-0.75-Temp-1-attention-3-layers-distil-with-6-layers-att-take-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 37.5364
• Wer: 0.3334

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: 0.002
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
43.7806	0.9	400	41.3073	0.2570
48.6549	1.8	800	41.8945	0.2740
57.4209	2.7	1200	39.9947	0.2872
68.8449	3.59	1600	39.4528	0.3059
79.4299	4.49	2000	38.9575	0.3179
93.0514	5.39	2400	37.5364	0.3334

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1
• Datasets 2.7.1
• Tokenizers 0.11.0