ssc-kcn-model

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: nan
• Cer: 0.9940
• Wer: 1.0

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
5.3972	0.1736	100	2.9985	0.9942	1.0
3.1473	0.3472	200	2.9632	0.9942	1.0
3.2868	0.5208	300	2.9640	0.9942	1.0
3.1277	0.6944	400	2.8740	0.9942	1.0
3.1015	0.8681	500	2.8538	0.9942	1.0
3.0851	1.0417	600	2.9601	0.9942	1.0
3.0431	1.2153	700	2.8703	0.9942	1.0
3.0662	1.3889	800	3.0285	0.9942	1.0
3.0358	1.5625	900	2.9641	0.9942	1.0
2.9999	1.7361	1000	3.0095	0.9942	1.0
3.0629	1.9097	1100	2.8662	0.9942	1.0
3.033	2.0833	1200	2.8406	0.9942	1.0
2.9814	2.2569	1300	2.8291	0.9942	1.0
3.0079	2.4306	1400	2.8602	0.9942	1.0
3.0814	2.6042	1500	2.8299	0.9942	1.0
3.0949	2.7778	1600	2.8245	0.9942	1.0
3.1026	2.9514	1700	2.8080	0.9942	1.0
3.0229	3.125	1800	2.9045	0.9942	1.0
3.0744	3.2986	1900	2.8780	0.9942	1.0
3.0559	3.4722	2000	2.9353	0.9942	1.0
3.0577	3.6458	2100	2.9453	0.9942	1.0
3.0844	3.8194	2200	3.0902	0.9942	1.0
3.2003	3.9931	2300	2.9683	0.9942	1.0
3.233	4.1667	2400	3.2298	0.9942	1.0
3.4188	4.3403	2500	3.3186	0.9942	1.0
3.5046	4.5139	2600	3.3847	0.9942	1.0
3.5387	4.6875	2700	3.4096	0.9942	1.0
3.517	4.8611	2800	3.4097	0.9942	1.0
3.5483	5.0347	2900	3.4097	0.9942	1.0
3.5612	5.2083	3000	3.4097	0.9942	1.0
3.5469	5.3819	3100	3.4097	0.9942	1.0
3.5421	5.5556	3200	3.4097	0.9942	1.0
3.5395	5.7292	3300	3.4097	0.9942	1.0
3.5186	5.9028	3400	3.4097	0.9942	1.0
3.5325	6.0764	3500	3.4097	0.9942	1.0
3.5384	6.25	3600	3.4097	0.9942	1.0
3.5326	6.4236	3700	3.4097	0.9942	1.0
3.513	6.5972	3800	3.4097	0.9942	1.0
3.5751	6.7708	3900	3.4097	0.9942	1.0
3.5294	6.9444	4000	3.4097	0.9942	1.0
3.545	7.1181	4100	3.4097	0.9942	1.0
3.5764	7.2917	4200	3.4097	0.9942	1.0
3.5258	7.4653	4300	3.4097	0.9942	1.0
3.514	7.6389	4400	3.4097	0.9942	1.0
3.5538	7.8125	4500	3.4097	0.9942	1.0
3.5231	7.9861	4600	3.4097	0.9942	1.0
3.5428	8.1597	4700	3.4097	0.9942	1.0
3.5614	8.3333	4800	3.4097	0.9942	1.0
3.5369	8.5069	4900	3.4097	0.9942	1.0
3.5457	8.6806	5000	3.4097	0.9942	1.0
3.5007	8.8542	5100	3.4097	0.9942	1.0
3.5424	9.0278	5200	3.4097	0.9942	1.0
3.5172	9.2014	5300	3.4097	0.9942	1.0
3.5408	9.375	5400	3.4097	0.9942	1.0
3.5382	9.5486	5500	3.4097	0.9942	1.0
3.5658	9.7222	5600	3.4097	0.9942	1.0
3.5203	9.8958	5700	3.4097	0.9942	1.0
3.5484	10.0694	5800	3.4097	0.9942	1.0
3.5278	10.2431	5900	3.4097	0.9942	1.0
3.5498	10.4167	6000	3.4097	0.9941	1.0
3.5557	10.5903	6100	3.4097	0.9942	1.0
3.5199	10.7639	6200	3.4097	0.9942	1.0
3.5358	10.9375	6300	3.4097	0.9942	1.0
3.5481	11.1111	6400	3.4097	0.9942	1.0
3.5652	11.2847	6500	3.4097	0.9942	1.0
3.5113	11.4583	6600	3.4097	0.9942	1.0
3.5728	11.6319	6700	3.4097	0.9942	1.0
3.512	11.8056	6800	3.4097	0.9942	1.0
3.5222	11.9792	6900	3.4097	0.9942	1.0
3.5793	12.1528	7000	3.4097	0.9942	1.0
3.5378	12.3264	7100	3.4097	0.9942	1.0
3.5463	12.5	7200	3.4097	0.9942	1.0
3.5221	12.6736	7300	3.4097	0.9942	1.0
3.5232	12.8472	7400	3.4097	0.9942	1.0
3.5345	13.0208	7500	3.4097	0.9942	1.0
3.545	13.1944	7600	3.4097	0.9942	1.0
3.5559	13.3681	7700	3.4097	0.9942	1.0
3.5081	13.5417	7800	3.4097	0.9942	1.0
3.5349	13.7153	7900	3.4097	0.9942	1.0
3.5193	13.8889	8000	3.4097	0.9942	1.0
3.5653	14.0625	8100	3.4097	0.9942	1.0
3.5493	14.2361	8200	3.4097	0.9942	1.0
3.525	14.4097	8300	3.4097	0.9942	1.0
3.5106	14.5833	8400	3.4097	0.9942	1.0
3.5627	14.7569	8500	3.4097	0.9942	1.0
3.5593	14.9306	8600	3.4097	0.9942	1.0
3.5159	15.1042	8700	3.4097	0.9941	1.0
3.5424	15.2778	8800	3.4097	0.9942	1.0
3.5829	15.4514	8900	3.4097	0.9942	1.0
3.5305	15.625	9000	3.4097	0.9942	1.0
3.5183	15.7986	9100	3.4097	0.9942	1.0
3.541	15.9722	9200	3.4097	0.9942	1.0
3.559	16.1458	9300	3.4097	0.9942	1.0
3.5281	16.3194	9400	3.4097	0.9942	1.0
3.5371	16.4931	9500	3.4097	0.9942	1.0
3.5843	16.6667	9600	3.4097	0.9942	1.0
3.5031	16.8403	9700	3.4097	0.9942	1.0
3.5339	17.0139	9800	3.4097	0.9942	1.0
3.5397	17.1875	9900	3.4097	0.9942	1.0
3.5408	17.3611	10000	3.4097	0.9942	1.0
3.5664	17.5347	10100	3.4097	0.9942	1.0
3.5145	17.7083	10200	3.4097	0.9942	1.0
3.5141	17.8819	10300	3.4097	0.9942	1.0
3.5562	18.0556	10400	3.4097	0.9942	1.0
3.5196	18.2292	10500	3.4097	0.9942	1.0
3.5335	18.4028	10600	3.4097	0.9942	1.0
3.5769	18.5764	10700	3.4097	0.9942	1.0
3.525	18.75	10800	3.4097	0.9942	1.0
3.5211	18.9236	10900	3.4097	0.9942	1.0
3.5441	19.0972	11000	3.4097	0.9942	1.0
3.5331	19.2708	11100	3.4097	0.9942	1.0
3.5274	19.4444	11200	3.4097	0.9942	1.0
3.5837	19.6181	11300	3.4097	0.9942	1.0
3.5116	19.7917	11400	3.4097	0.9942	1.0
3.5495	19.9653	11500	3.4097	0.9942	1.0
6.9849	20.1389	11600	nan	0.9940	1.0
0.0	20.3125	11700	nan	0.9940	1.0
0.0	20.4861	11800	nan	0.9940	1.0
0.0	20.6597	11900	nan	0.9940	1.0
0.0	20.8333	12000	nan	0.9940	1.0
0.0	21.0069	12100	nan	0.9940	1.0
0.0	21.1806	12200	nan	0.9940	1.0
0.0	21.3542	12300	nan	0.9940	1.0
0.0	21.5278	12400	nan	0.9940	1.0
0.0	21.7014	12500	nan	0.9940	1.0
0.0	21.875	12600	nan	0.9940	1.0
0.0	22.0486	12700	nan	0.9940	1.0
0.0	22.2222	12800	nan	0.9940	1.0
0.0	22.3958	12900	nan	0.9940	1.0
0.0	22.5694	13000	nan	0.9940	1.0
0.0	22.7431	13100	nan	0.9940	1.0
0.0	22.9167	13200	nan	0.9940	1.0
0.0	23.0903	13300	nan	0.9940	1.0
0.0	23.2639	13400	nan	0.9940	1.0
0.0	23.4375	13500	nan	0.9940	1.0
0.0	23.6111	13600	nan	0.9940	1.0
0.0	23.7847	13700	nan	0.9940	1.0
0.0	23.9583	13800	nan	0.9940	1.0
0.0	24.1319	13900	nan	0.9940	1.0
0.0	24.3056	14000	nan	0.9940	1.0
0.0	24.4792	14100	nan	0.9940	1.0
0.0	24.6528	14200	nan	0.9940	1.0
0.0	24.8264	14300	nan	0.9940	1.0
0.0	25.0	14400	nan	0.9940	1.0
0.0	25.1736	14500	nan	0.9940	1.0
0.0	25.3472	14600	nan	0.9940	1.0
0.0	25.5208	14700	nan	0.9940	1.0
0.0	25.6944	14800	nan	0.9940	1.0
0.0	25.8681	14900	nan	0.9940	1.0
0.0	26.0417	15000	nan	0.9940	1.0
0.0	26.2153	15100	nan	0.9940	1.0
0.0	26.3889	15200	nan	0.9940	1.0
0.0	26.5625	15300	nan	0.9940	1.0
0.0	26.7361	15400	nan	0.9940	1.0
0.0	26.9097	15500	nan	0.9940	1.0
0.0	27.0833	15600	nan	0.9940	1.0
0.0	27.2569	15700	nan	0.9940	1.0
0.0	27.4306	15800	nan	0.9940	1.0
0.0	27.6042	15900	nan	0.9940	1.0
0.0	27.7778	16000	nan	0.9940	1.0
0.0	27.9514	16100	nan	0.9940	1.0
0.0	28.125	16200	nan	0.9940	1.0
0.0	28.2986	16300	nan	0.9940	1.0
0.0	28.4722	16400	nan	0.9940	1.0
0.0	28.6458	16500	nan	0.9940	1.0
0.0	28.8194	16600	nan	0.9940	1.0
0.0	28.9931	16700	nan	0.9940	1.0
0.0	29.1667	16800	nan	0.9940	1.0
0.0	29.3403	16900	nan	0.9940	1.0
0.0	29.5139	17000	nan	0.9940	1.0
0.0	29.6875	17100	nan	0.9940	1.0
0.0	29.8611	17200	nan	0.9940	1.0

Framework versions

• Transformers 4.57.2
• Pytorch 2.9.1+cu128
• Datasets 3.6.0
• Tokenizers 0.22.0