license_Plate_Recognizer

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

• Loss: 0.3905
• Cer: 0.0305

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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
0.8435	1.0	128	0.5212	0.0798
0.326	2.0	256	0.4098	0.0609
0.2681	3.0	384	0.5235	0.0793
0.2984	4.0	512	0.3912	0.0590
0.1671	5.0	640	0.3756	0.0461
0.1827	6.0	768	0.3875	0.0518
0.1368	7.0	896	0.3851	0.0566
0.1304	8.0	1024	0.3917	0.0540
0.148	9.0	1152	0.4448	0.0648
0.108	10.0	1280	0.3771	0.0441
0.0756	11.0	1408	0.3486	0.0428
0.0891	12.0	1536	0.3960	0.0504
0.0832	13.0	1664	0.3739	0.0438
0.0685	14.0	1792	0.3805	0.0428
0.0642	15.0	1920	0.3541	0.0416
0.0621	16.0	2048	0.3981	0.0457
0.0335	17.0	2176	0.4031	0.0447
0.0376	18.0	2304	0.4305	0.0520
0.0534	19.0	2432	0.4086	0.0432
0.0314	20.0	2560	0.4166	0.0408
0.0219	21.0	2688	0.4393	0.0409
0.0425	22.0	2816	0.4522	0.0476
0.0186	23.0	2944	0.4166	0.0395
0.0311	24.0	3072	0.3867	0.0373
0.0191	25.0	3200	0.3832	0.0411
0.0143	26.0	3328	0.3954	0.0382
0.0201	27.0	3456	0.3959	0.0388
0.0196	28.0	3584	0.4099	0.0371
0.014	29.0	3712	0.4205	0.0388
0.0079	30.0	3840	0.4231	0.0381
0.0082	31.0	3968	0.4497	0.0400
0.0067	32.0	4096	0.4340	0.0361
0.0191	33.0	4224	0.4181	0.0352
0.0059	34.0	4352	0.4159	0.0351
0.0049	35.0	4480	0.4076	0.0328
0.0073	36.0	4608	0.4093	0.0333
0.0025	37.0	4736	0.3903	0.0335
0.0056	38.0	4864	0.4182	0.0370
0.0029	39.0	4992	0.3996	0.0332
0.0012	40.0	5120	0.3948	0.0325
0.0041	41.0	5248	0.4099	0.0315
0.0048	42.0	5376	0.4057	0.0319
0.0016	43.0	5504	0.4029	0.0316
0.0012	44.0	5632	0.3905	0.0305
0.0015	45.0	5760	0.4043	0.0310
0.0014	46.0	5888	0.3977	0.0310
0.0028	47.0	6016	0.4055	0.0317
0.0012	48.0	6144	0.4048	0.0311
0.0048	49.0	6272	0.4018	0.0305
0.0009	50.0	6400	0.4035	0.0306

Framework versions

• Transformers 4.44.2
• Pytorch 2.6.0+cu124
• Datasets 3.0.1
• Tokenizers 0.19.1