mms-1b-all-lin-Fleurs_AMMI_AFRIVOICE_LRSC-20hrs-v1

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 0.001
train_batch_size: 8
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Use OptimizerNames.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: 100
num_epochs: 100
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.9973	1.0	356	0.3579	0.2562	0.0743
0.4557	2.0	712	0.3197	0.2418	0.0700
0.4169	3.0	1068	0.3057	0.2421	0.0695
0.398	4.0	1424	0.3011	0.2487	0.0713
0.37	5.0	1780	0.2914	0.2358	0.0675
0.3584	6.0	2136	0.2890	0.2061	0.0625
0.3502	7.0	2492	0.2792	0.2210	0.0651
0.3399	8.0	2848	0.2770	0.2100	0.0614
0.3335	9.0	3204	0.2815	0.2162	0.0637
0.3308	10.0	3560	0.2733	0.2106	0.0625
0.3238	11.0	3916	0.2730	0.2171	0.0627
0.3077	12.0	4272	0.2776	0.2046	0.0607
0.3092	13.0	4628	0.2667	0.1980	0.0590
0.3043	14.0	4984	0.2901	0.2089	0.0621
0.3006	15.0	5340	0.2687	0.1986	0.0582
0.2973	16.0	5696	0.2812	0.2041	0.0612
0.2847	17.0	6052	0.2855	0.1941	0.0580
0.2823	18.0	6408	0.2804	0.2035	0.0603
0.2772	19.0	6764	0.2734	0.2129	0.0625
0.2723	20.0	7120	0.2799	0.2100	0.0635
0.2671	21.0	7476	0.2702	0.2323	0.0682
0.2622	22.0	7832	0.2745	0.2086	0.0636
0.2616	23.0	8188	0.2820	0.2045	0.0620
0.252	24.0	8544	0.2819	0.2342	0.0761
0.2531	25.0	8900	0.2806	0.1954	0.0586
0.249	26.0	9256	0.2792	0.1979	0.0596
0.2431	27.0	9612	0.2919	0.1950	0.0589