Skip to content

Training the OCR Model

The train script launches the end-to-end training process for a license plate OCR model, using:

  • A user-defined model architecture (model_config.yaml)
  • A plate configuration (plate_config.yaml)
  • A training/validation dataset in CSV format
  • Optional augmentation, loss, optimizer and callback settings

Basic Usage

# You can set the backend to either TensorFlow, JAX or PyTorch
KERAS_BACKEND=tensorflow fast-plate-ocr train \
  --model-config-file models/cct_s_v1.yaml \
  --plate-config-file config/latin_plates.yaml \
  --annotations data/train.csv \
  --val-annotations data/val.csv \
  --epochs 150 \
  --batch-size 64 \
  --output-dir trained_models/
Keras Backend

Make sure you have installed a backend for training. See Backend section for more details.

Tip

Use --weights-path <path/to/model.keras> when fine-tuning. All pre-trained models are uploaded to this GH release > Assets.

Logging and Checkpoints

The script saves:

  • Best model (based on monitored metric)
  • Final model weights
  • Training parameters and configs
  • Optional TensorBoard logs

Metrics Tracked

The model is compiled with:

  • plate_acc: Fully correct plate predictions
  • cat_acc: Character-level accuracy
  • top_3_k: Accuracy if target is in top-3 predictions
  • plate_len_acc: Plate length matches

See Metrics section for more details about metrics.

View all CLI flags

The CLI supports over 30 options. You can view them with:

fast-plate-ocr train --help