learning_rate: 0.0003
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 64
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: 400
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer
0.4001	1.0	29	0.3761	0.006	0.1412
0.2049	2.0	58	0.3401	0.006	0.1411
0.1038	3.0	87	0.3867	0.006	0.1323
0.0611	4.0	116	0.4391	0.006	0.1493
0.0577	5.0	145	0.4881	0.006	0.1542
0.0631	6.0	174	0.5029	0.006	0.1607
0.068	7.0	203	0.5696	0.006	0.1721
0.0788	8.0	232	0.5848	0.006	0.1783
0.0814	9.0	261	0.6760	0.006	0.1807
0.1012	9.6609	280	0.6575	0.006	0.1862

mdsingh2024/ap-lWoqAtb6o7NMEmtpZzXNvv

ap-lWoqAtb6o7NMEmtpZzXNvv