ssc-cgg-mms-model-mix-adapt-max2

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7049
• Cer: 0.1372
• Wer: 0.6029

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: 0.001
• train_batch_size: 1
• eval_batch_size: 6
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 2
• 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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
0.4633	0.2261	200	0.7482	0.1501	0.6867
0.488	0.4522	400	0.7169	0.1465	0.6558
0.4615	0.6783	600	0.7037	0.1422	0.6081
0.408	0.9045	800	0.7335	0.1471	0.6367
0.4259	1.1300	1000	0.7168	0.1477	0.6399
0.4754	1.3561	1200	0.7328	0.1418	0.6101
0.4259	1.5822	1400	0.7273	0.1429	0.6249
0.4098	1.8084	1600	0.7386	0.1454	0.6378
0.442	2.0339	1800	0.7131	0.1462	0.6346
0.4213	2.2600	2000	0.7259	0.1481	0.6367
0.4142	2.4862	2200	0.7202	0.1423	0.6228
0.4241	2.7123	2400	0.7311	0.1429	0.6282
0.3955	2.9384	2600	0.7247	0.1437	0.6244
0.4146	3.1639	2800	0.7132	0.1466	0.6429
0.4079	3.3901	3000	0.7245	0.1408	0.6122
0.41	3.6162	3200	0.7120	0.1434	0.6163
0.3987	3.8423	3400	0.7154	0.1426	0.6269
0.4015	4.0678	3600	0.7122	0.1354	0.5907
0.3837	4.2940	3800	0.7201	0.1384	0.5944
0.3412	4.5201	4000	0.7029	0.1397	0.6171
0.373	4.7462	4200	0.7318	0.1414	0.6292
0.3778	4.9723	4400	0.7155	0.1391	0.6166
0.3393	5.1979	4600	0.7126	0.1393	0.6213
0.3146	5.4240	4800	0.7337	0.1407	0.6201
0.3224	5.6501	5000	0.7114	0.1391	0.6126
0.3313	5.8762	5200	0.7224	0.1388	0.6062
0.3166	6.1018	5400	0.7191	0.1394	0.6191
0.3217	6.3279	5600	0.7007	0.1370	0.6041
0.3888	6.5540	5800	0.7155	0.1356	0.5895
0.3669	6.7801	6000	0.7153	0.1396	0.6172
0.3219	7.0057	6200	0.7100	0.1375	0.6066
0.3403	7.2318	6400	0.7135	0.1392	0.6076
0.3155	7.4579	6600	0.7133	0.1389	0.6112
0.3226	7.6840	6800	0.7107	0.1382	0.6107
0.3742	7.9101	7000	0.7112	0.1379	0.6038
0.3064	8.1357	7200	0.7102	0.1378	0.6060
0.3204	8.3618	7400	0.7083	0.1371	0.6088
0.2809	8.5879	7600	0.7119	0.1382	0.6116
0.2899	8.8140	7800	0.7077	0.1379	0.6079
0.3445	9.0396	8000	0.7082	0.1376	0.6066
0.3208	9.2657	8200	0.7039	0.1382	0.6053
0.3093	9.4918	8400	0.7087	0.1372	0.6048
0.3084	9.7179	8600	0.7098	0.1370	0.6046
0.3444	9.9440	8800	0.7049	0.1372	0.6029

Framework versions

• Transformers 4.52.1
• Pytorch 2.9.1+cu128
• Datasets 3.6.0
• Tokenizers 0.21.4