learning_rate: 0.0001
train_batch_size: 8
eval_batch_size: 8
seed: 42
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: 1000
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
8.6728	1.3055	500	3.0311	1.0004
2.1354	2.6110	1000	1.3591	0.9853
1.1544	3.9164	1500	0.9539	0.8498
0.7780	5.2219	2000	0.7803	0.7699
0.5782	6.5274	2500	0.8055	0.7366
0.4490	7.8329	3000	0.6757	0.6613
0.3674	9.1384	3500	0.6647	0.6290
0.2910	10.4439	4000	0.6616	0.6186
0.2483	11.7493	4500	0.6115	0.6122
0.2110	13.0548	5000	0.6493	0.6075
0.1764	14.3603	5500	0.6653	0.5875
0.1573	15.6658	6000	0.6814	0.5857
0.1375	16.9713	6500	0.6300	0.5559
0.1254	18.2768	7000	0.6974	0.5495
0.1085	19.5822	7500	0.6578	0.5559
0.1009	20.8877	8000	0.6518	0.5391
0.0791	22.1932	8500	0.6833	0.5305
0.0689	23.4987	9000	0.7359	0.5269
0.0595	24.8042	9500	0.7032	0.5115
0.0707	26.1097	10000	0.6596	0.5054
0.0486	27.4151	10500	0.6564	0.5104
0.0524	28.7206	11000	0.6712	0.5054
0.0524	30.0	11490	0.6853	0.5039

Iram076/wav2vec2-english-learner-google-colab

wav2vec2-english-learner-google-colab