ASR_Germen_Wav2Vec2_Local_Train

This model is a fine-tuned version of facebook/wav2vec2-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3619
• Per: 0.1676

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 300
• training_steps: 15000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per
0.6277	0.1630	3000	0.5961	0.2501
0.445	0.3259	6000	0.4649	0.2070
0.3751	0.4889	9000	0.4084	0.1818
0.4281	0.6518	12000	0.3907	0.1792
0.3835	0.8148	15000	0.3619	0.1676

Framework versions

• Transformers 4.57.0
• Pytorch 2.9.0+cu130
• Datasets 2.18.0
• Tokenizers 0.22.1