learning_rate: 2e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
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_ratio: 0.2
training_steps: 1000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
0.0221	25.0	100	0.7679	28.9474
0.0048	50.0	200	0.7079	23.6842
0.0042	75.0	300	0.8937	21.0526
0.0	100.0	400	0.9432	21.0526
0.0	125.0	500	0.9563	21.0526
0.0	150.0	600	0.9644	21.0526
0.0	175.0	700	0.9711	21.0526
0.0	200.0	800	0.9751	21.0526
0.0	225.0	900	0.9771	21.0526
0.0	250.0	1000	0.9780	21.0526

CSY1109/drone_sy_tiny_t4

Drone test En - Siang Yi