learning_rate: 0.0001
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
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: 200
num_epochs: 15
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Accuracy
0.0282	2.2727	100	0.0080	0.2	97.44
0.0019	4.5455	200	0.0044	0.15	98.08
0.0008	6.8182	300	0.0041	0.15	98.08
0.0006	9.0909	400	0.0040	0.1	98.72
0.0004	11.3636	500	0.0040	0.1	98.72
0.0004	13.6364	600	0.0040	0.1	98.72

Iremozcelik/pt_transcriber_1.0

pt_transcriber_1.0