learning_rate: 1e-05
train_batch_size: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 16
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: 50
num_epochs: 5

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.4184	0.5674	10	1.0391	0.3636	0.1290
1.308	1.1135	20	1.0367	0.3409	0.1210
1.375	1.6809	30	1.0398	0.2727	0.0968
1.3359	2.2270	40	1.0344	0.3636	0.1290
1.352	2.7943	50	1.0320	0.3409	0.1210
1.366	3.3404	60	1.0297	0.3636	0.1290
1.3701	3.9078	70	1.0250	0.2727	0.0968
1.2679	4.4539	80	1.0266	0.3409	0.1210

Jeff13680/whisper-lora-642

whisper-lora-642