learning_rate: 0.001
train_batch_size: 2
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: 100
num_epochs: 5
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
7.4923	0.1546	100	4.8960	1.0122
4.5397	0.3091	200	4.2869	0.9985
4.2989	0.4637	300	4.2238	0.9997
3.3463	0.6182	400	3.9085	0.9997
3.8475	0.7728	500	3.6456	0.9977
3.2022	0.9274	600	2.5560	0.9990
1.9604	1.0819	700	1.4278	0.9718
1.3657	1.2365	800	1.1949	0.9418
1.2215	1.3910	900	1.1053	0.8904
1.1283	1.5456	1000	1.1093	0.9243
1.1446	1.7002	1100	1.0480	0.8440
1.0842	1.8547	1200	1.0411	0.8180
1.0543	2.0093	1300	1.0096	0.8379
1.0568	2.1638	1400	1.0144	0.7895
0.9882	2.3184	1500	0.9860	0.7840
1.0208	2.4730	1600	0.9688	0.7756
0.9466	2.6275	1700	0.9622	0.8350
0.9941	2.7821	1800	0.9579	0.7489
1.0361	2.9366	1900	0.9332	0.7598
1.0198	3.0912	2000	0.9150	0.7420
0.9461	3.2457	2100	0.9212	0.7660
0.8931	3.4003	2200	0.9153	0.7302
0.9039	3.5549	2300	0.9111	0.7295
0.9223	3.7094	2400	0.8941	0.7354
0.8911	3.8640	2500	0.8955	0.7103
0.912	4.0185	2600	0.8821	0.7065
0.8834	4.1731	2700	0.8928	0.7073
0.9154	4.3277	2800	0.8806	0.7124
0.914	4.4822	2900	0.8843	0.7007
0.9183	4.6368	3000	0.8807	0.6961
0.7955	4.7913	3100	0.8763	0.6939
0.8668	4.9459	3200	0.8735	0.6953

robertp408/wav2vec2-large-mms-1b-aft-rwm

wav2vec2-large-mms-1b-aft-rwm