wav2vec2-gcf-r8

This model is a fine-tuned version of LLL-CREAM/wav2vec2-HAT-0.2K-ALH-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.9370
• Wer: 96.07

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 100
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
7.4949	2.3474	1000	3.9160	96.07
7.4919	4.6948	2000	3.9299	96.07
7.3981	7.0423	3000	3.9875	96.07
7.4559	9.3897	4000	3.9690	96.07
7.4886	11.7371	5000	4.0359	96.07
7.4874	14.0845	6000	3.9979	96.07
7.4587	16.4319	7000	3.9472	96.07
7.4759	18.7793	8000	3.9370	96.07

Framework versions

• Transformers 5.5.0
• Pytorch 2.4.1+cu124
• Datasets 3.6.0
• Tokenizers 0.22.2