learning_rate: 0.0001
train_batch_size: 8
eval_batch_size: 8
seed: 42
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: 50
num_epochs: 3
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Cer
44.6688	0.1289	200	5.3220	100.0
5.0381	0.2579	400	4.6642	100.0
4.8478	0.3868	600	4.6375	100.0
4.835	0.5158	800	4.6517	100.0
4.7158	0.6447	1000	4.6317	100.0
4.48	0.7737	1200	4.2848	99.9354
3.5034	0.9026	1400	2.8766	51.6627
2.7033	1.0316	1600	2.4780	45.2644
2.3513	1.1605	1800	2.2651	42.3913
2.0186	1.2895	2000	1.9799	36.8096
1.8504	1.4184	2200	1.9017	35.7873
1.6456	1.5474	2400	1.6934	32.8555
1.5546	1.6763	2600	1.5812	31.0576
1.4543	1.8053	2800	1.4790	29.9354
1.3236	1.9342	3000	1.4232	29.2949
1.2239	2.0632	3200	1.3445	27.3678
1.0797	2.1921	3400	1.2819	26.4219
1.0512	2.3211	3600	1.2911	27.0564
0.9936	2.4500	3800	1.1997	25.4994
0.9784	2.5790	4000	1.2025	25.6522
0.9561	2.7079	4200	1.1842	24.8179
0.9128	2.8369	4400	1.1630	24.8237
0.886	2.9658	4600	1.1541	24.7591

Gummybear05/wav2vec2-E10_freq_speed2

wav2vec2-E10_freq_speed2