ultravox-salt-asr

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

• Loss: 1.9282

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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: cosine
• lr_scheduler_warmup_steps: 500
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
15.9454	0.1665	25	2.0889
15.7323	0.3331	50	2.0713
15.3988	0.4996	75	2.0468
15.199	0.6661	100	2.0189
15.1278	0.8326	125	2.0081
14.9351	0.9992	150	1.9898
14.1657	1.1599	175	1.9887
14.7294	1.3264	200	1.9699
14.8916	1.4929	225	1.9669
13.9667	1.6595	250	1.9196
14.2969	1.8260	275	1.9168
14.091	1.9925	300	1.9282

Framework versions

• Transformers 4.54.0
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4