w2v-bert-2.0-Fleurs_AMMI_AFRIVOICE_LRSC-ln-10hrs-v1

This model is a fine-tuned version of facebook/w2v-bert-2.0 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: 5e-05
train_batch_size: 8
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
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
num_epochs: 100
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.4981	1.0	192	2.8671	0.9999	0.9897
1.6608	2.0	384	0.6305	0.4155	0.1240
0.5157	3.0	576	0.5014	0.3768	0.1096
0.3888	4.0	768	0.4831	0.3304	0.1022
0.3245	5.0	960	0.4152	0.2813	0.0879
0.2763	6.0	1152	0.4413	0.2688	0.0851
0.2328	7.0	1344	0.4405	0.2748	0.0852
0.2014	8.0	1536	0.4643	0.2628	0.0800
0.1733	9.0	1728	0.4771	0.2805	0.0847
0.1544	10.0	1920	0.4879	0.2505	0.0784
0.1318	11.0	2112	0.5000	0.2383	0.0760
0.1266	12.0	2304	0.5216	0.2546	0.0782
0.1086	13.0	2496	0.5079	0.2431	0.0736
0.0938	14.0	2688	0.5661	0.2323	0.0719
0.0806	15.0	2880	0.5603	0.2272	0.0710
0.0721	16.0	3072	0.5914	0.2279	0.0728
0.0665	17.0	3264	0.6192	0.2261	0.0725
0.062	18.0	3456	0.5625	0.2583	0.0787
0.0556	19.0	3648	0.5995	0.2371	0.0773
0.0487	20.0	3840	0.5746	0.2464	0.0744
0.0401	21.0	4032	0.6108	0.2469	0.0737
0.0382	22.0	4224	0.6207	0.2397	0.0774
0.035	23.0	4416	0.6793	0.2461	0.0752
0.0302	24.0	4608	0.5576	0.2404	0.0735
0.0244	25.0	4800	0.6412	0.2190	0.0704
0.022	26.0	4992	0.6204	0.2355	0.0753
0.0225	27.0	5184	0.6687	0.2248	0.0706
0.0172	28.0	5376	0.6542	0.2355	0.0749
0.0166	29.0	5568	0.6507	0.2242	0.0713
0.0131	30.0	5760	0.6806	0.2292	0.0713
0.0121	31.0	5952	0.6924	0.2377	0.0749
0.0115	32.0	6144	0.6961	0.2283	0.0732
0.011	33.0	6336	0.7073	0.2269	0.0712
0.0096	34.0	6528	0.7129	0.2325	0.0714
0.0086	35.0	6720	0.7263	0.2538	0.0790