File size: 1,765 Bytes
f8584da 817b858 f8584da 11d25a2 f8584da 11d25a2 f8584da 817b858 f8584da 11d25a2 f8584da 817b858 f8584da 11d25a2 f8584da 817b858 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 |
---
tags:
- generated_from_trainer
model-index:
- name: trocr-base-printed_license_plates_ocr
results: []
language:
- en
metrics:
- cer
pipeline_tag: image-to-text
---
# trocr-base-printed_license_plates_ocr
This model is a fine-tuned version of [microsoft/trocr-base-printed](https://huggingface.co/microsoft/trocr-base-printed).
It achieves the following results on the evaluation set:
- Loss: 0.1581
- CER: 0.0368
## Model description
This model extracts text from image input (License Plates).
For more information on how it was created, check out the following link: https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/blob/main/Optical%20Character%20Recognition%20(OCR)/OCR%20License%20Plates/OCR_license_plate_text_recognition.ipynb
## Intended uses & limitations
This model is intended to demonstrate my ability to solve a complex problem using technology. You are welcome to test and experiment with this model, but it is at your own risk/peril.
## Training and evaluation data
Dataset Source: https://www.kaggle.com/datasets/nickyazdani/license-plate-text-recognition-dataset
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | CER |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.3144 | 1.0 | 2000 | 0.2463 | 0.0473 |
| 0.143 | 2.0 | 4000 | 0.1581 | 0.0368 |
### Framework versions
- Transformers 4.21.3
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1 |