WAXAL-ful_asr-w2v-bert-2.0

This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.0337
• Cer: 0.2289

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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: 500
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
3.9181	1.7881	300	0.9619	0.2445
0.9746	3.5731	600	0.9351	0.2335
0.7613	5.3582	900	0.9488	0.2282
0.5810	7.1433	1200	0.9767	0.2317
0.4433	8.9313	1500	0.9938	0.2281
0.4433	10.0	1680	1.0337	0.2289

Framework versions

• Transformers 5.12.1
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2