learning_rate: 1e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
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: 1000
num_epochs: 1
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
0.3419	0.0969	3000	0.5707	47.4670
0.279	0.1938	6000	0.4981	30.5959
0.255	0.2907	9000	0.4742	39.2586
0.24	0.3876	12000	0.4524	33.4547
0.2279	0.4845	15000	0.4341	31.1897
0.2172	0.5814	18000	0.4281	33.8020
0.2136	0.6783	21000	0.4176	31.4035
0.2088	0.7751	24000	0.4162	31.2928
0.2047	0.8720	27000	0.4140	32.2558
0.2	0.9689	30000	0.4065	31.7637

mondhs/l3-whisper-small-l3c2_e1_v3

l3-whisper-small-l3c2_e1_v3