wav2vec2-ft-lre6-adm-ga1b16

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

• Loss: 0.9080
• Wer: 1.0
• Cer: 1.0

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: 1e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use 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: 500
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.211	0.2082	1000	1.0226	1.0	1.0
1.1194	0.4165	2000	0.9400	1.0	1.0
0.8982	0.6247	3000	0.8997	1.0	1.0
0.8637	0.8330	4000	0.9401	1.0	1.0
0.8279	1.0412	5000	0.9447	1.0	1.0
0.8102	1.2495	6000	0.9113	1.0	1.0
0.8467	1.4577	7000	0.9032	1.0	1.0
0.8405	1.6660	8000	0.8879	1.0	1.0
0.9392	1.8742	9000	0.8960	1.0	1.0
0.8377	2.0825	10000	0.9080	1.0	1.0

Framework versions

• Transformers 4.52.3
• Pytorch 2.5.1
• Datasets 3.6.0
• Tokenizers 0.21.1