license_plate_recognizer

This model is a fine-tuned version of microsoft/trocr-base-printed on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0110
• Cer: 0.0044

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.2458	1.0	199	0.0687	0.0102
0.1974	2.0	398	0.0810	0.0120
0.0876	3.0	597	0.0436	0.0049
0.0529	4.0	796	0.0295	0.0035
0.0432	5.0	995	0.0291	0.0053
0.0247	6.0	1194	0.0202	0.0040
0.0367	7.0	1393	0.0196	0.0049
0.0349	8.0	1592	0.0165	0.0049
0.0144	9.0	1791	0.0242	0.0044
0.0155	10.0	1990	0.0119	0.0018
0.034	11.0	2189	0.0099	0.0035
0.0135	12.0	2388	0.0142	0.0040
0.0188	13.0	2587	0.0156	0.0053
0.0083	14.0	2786	0.0193	0.0035
0.0092	15.0	2985	0.0097	0.0031
0.0131	16.0	3184	0.0125	0.0018
0.0074	17.0	3383	0.0149	0.0031
0.0028	18.0	3582	0.0198	0.0044
0.009	19.0	3781	0.0158	0.0058
0.0035	20.0	3980	0.0091	0.0031
0.0083	21.0	4179	0.0102	0.0040
0.0035	22.0	4378	0.0060	0.0018
0.0059	23.0	4577	0.0132	0.0031
0.0022	24.0	4776	0.0094	0.0027
0.0093	25.0	4975	0.0143	0.0027
0.0007	26.0	5174	0.0145	0.0035
0.0004	27.0	5373	0.0116	0.0035
0.0029	28.0	5572	0.0148	0.0035
0.006	29.0	5771	0.0131	0.0031
0.0002	30.0	5970	0.0111	0.0027
0.0002	31.0	6169	0.0102	0.0031
0.0004	32.0	6368	0.0093	0.0022
0.0001	33.0	6567	0.0176	0.0040
0.0001	34.0	6766	0.0103	0.0040
0.0001	35.0	6965	0.0097	0.0027
0.0036	36.0	7164	0.0103	0.0031
0.0002	37.0	7363	0.0101	0.0031
0.0006	38.0	7562	0.0109	0.0027
0.0005	39.0	7761	0.0163	0.0031
0.0001	40.0	7960	0.0142	0.0027
0.0002	41.0	8159	0.0104	0.0022
0.0001	42.0	8358	0.0110	0.0035
0.0	43.0	8557	0.0106	0.0035
0.0	44.0	8756	0.0104	0.0031
0.0006	45.0	8955	0.0102	0.0031
0.0001	46.0	9154	0.0104	0.0031
0.0	47.0	9353	0.0103	0.0031
0.0	48.0	9552	0.0110	0.0031
0.0	49.0	9751	0.0111	0.0031
0.0	50.0	9950	0.0111	0.0031

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1