xlsr-nomi-nmcpc

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0000
• Wer: 0.2574

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0004
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 132
• num_epochs: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.7569	3.8835	200	3.0327	1.0
2.9033	7.7670	400	2.5777	1.0
2.006	11.6505	600	0.7689	0.7298
0.7612	15.5340	800	0.2026	0.4191
0.3581	19.4175	1000	0.0966	0.3489
0.2432	23.3010	1200	0.0406	0.3128
0.1625	27.1845	1400	0.0479	0.3106
0.1391	31.0680	1600	0.0172	0.2915
0.1157	34.9515	1800	0.0110	0.2851
0.1016	38.8350	2000	0.0096	0.2766
0.0807	42.7184	2200	0.0162	0.2723
0.0673	46.6019	2400	0.0093	0.2681
0.0695	50.4854	2600	0.0305	0.2851
0.0589	54.3689	2800	0.0058	0.2638
0.0508	58.2524	3000	0.0172	0.2660
0.0503	62.1359	3200	0.0070	0.2660
0.0516	66.0194	3400	0.0017	0.2660
0.0435	69.9029	3600	0.0037	0.2617
0.0378	73.7864	3800	0.0020	0.2638
0.0325	77.6699	4000	0.0061	0.2702
0.031	81.5534	4200	0.0036	0.2617
0.0292	85.4369	4400	0.0075	0.2660
0.0281	89.3204	4600	0.0006	0.2532
0.0334	93.2039	4800	0.0007	0.2574
0.0307	97.0874	5000	0.0029	0.2596
0.0232	100.9709	5200	0.0025	0.2553
0.0228	104.8544	5400	0.0010	0.2596
0.0216	108.7379	5600	0.0014	0.2638
0.0207	112.6214	5800	0.0005	0.2596
0.0163	116.5049	6000	0.0021	0.2596
0.0232	120.3883	6200	0.0173	0.2638
0.0247	124.2718	6400	0.0102	0.2596
0.0198	128.1553	6600	0.0003	0.2553
0.0124	132.0388	6800	0.0001	0.2574
0.0117	135.9223	7000	0.0001	0.2574
0.0131	139.8058	7200	0.0002	0.2574
0.0119	143.6893	7400	0.0001	0.2574
0.0103	147.5728	7600	0.0001	0.2574
0.0083	151.4563	7800	0.0001	0.2574
0.0115	155.3398	8000	0.0001	0.2574
0.0112	159.2233	8200	0.0001	0.2574
0.0128	163.1068	8400	0.0001	0.2574
0.0062	166.9903	8600	0.0001	0.2574
0.0057	170.8738	8800	0.0000	0.2574
0.0072	174.7573	9000	0.0000	0.2574
0.0061	178.6408	9200	0.0000	0.2574
0.0047	182.5243	9400	0.0000	0.2574
0.0063	186.4078	9600	0.0000	0.2574
0.0043	190.2913	9800	0.0000	0.2574
0.006	194.1748	10000	0.0000	0.2574
0.0042	198.0583	10200	0.0000	0.2574

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1