fineTunning3

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.7642
• Wer: 1.0100

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	4.03	250	12.8710	1.0
7.1272	8.06	500	3.4483	1.0
7.1272	12.1	750	3.0834	1.0
2.9944	16.13	1000	3.0052	1.0
2.9944	20.16	1250	2.8564	1.0
2.7966	24.19	1500	2.6613	1.0
2.7966	28.22	1750	2.3765	1.0
2.4567	32.26	2000	2.0220	1.0089
2.4567	36.29	2250	1.7642	1.0100

Framework versions

• Transformers 4.17.0
• Pytorch 2.5.1+cu118
• Datasets 1.18.3
• Tokenizers 0.20.3