wav2vec2-xlsr53-algerian-east-v2

This model is a fine-tuned version of khier12/800min_wav2vec_xls-r53_FT_Algerian_Dialect 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: 0.0001
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 32
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: 500
num_epochs: 15
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
40.1153	0.3655	100	1.6791	0.7990
26.2294	0.7311	200	1.1508	0.7318
21.9277	1.0950	300	0.9950	0.6864
20.5062	1.4606	400	0.9540	0.6536
19.63	1.8261	500	0.8934	0.6299
18.2782	2.1901	600	0.8425	0.6117
17.5983	2.5556	700	0.8034	0.5935
17.1768	2.9212	800	0.7940	0.5833
16.5101	3.2851	900	0.7743	0.5732
16.0058	3.6507	1000	0.7721	0.5571
15.3478	4.0146	1100	0.7812	0.5528
15.0882	4.3802	1200	0.7235	0.5365
14.7465	4.7457	1300	0.7228	0.5292
14.6741	5.1097	1400	0.7418	0.5251
14.0926	5.4752	1500	0.7005	0.5197
14.3768	5.8408	1600	0.6987	0.5217
13.3279	6.2047	1700	0.7156	0.5131
13.5695	6.5703	1800	0.6975	0.5156
13.3755	6.9358	1900	0.6949	0.5070
12.9600	7.2997	2000	0.6775	0.4995
12.9266	7.6653	2100	0.6770	0.5029
12.7995	8.0292	2200	0.6826	0.5023
12.4408	8.3948	2300	0.6657	0.4911
12.6026	8.7603	2400	0.6670	0.4885
12.5762	9.1243	2500	0.6804	0.4873
12.0991	9.4898	2600	0.6679	0.4887
11.9747	9.8554	2700	0.6583	0.4770
11.6910	10.2193	2800	0.6745	0.4741
11.8903	10.5849	2900	0.6674	0.4791
11.6509	10.9504	3000	0.6622	0.4729
11.9245	11.3144	3100	0.6536	0.4734
11.2273	11.6799	3200	0.6676	0.4761
11.4728	12.0439	3300	0.6535	0.4720
11.3458	12.4094	3400	0.6503	0.4708
10.9733	12.7750	3500	0.6401	0.4693
11.2605	13.1389	3600	0.6574	0.4696
11.1309	13.5045	3700	0.6443	0.4671
11.2028	13.8700	3800	0.6490	0.4658
10.8542	14.2340	3900	0.6486	0.4658
10.9644	14.5995	4000	0.6495	0.4646
11.0140	14.9650	4100	0.6473	0.4637