---
license: apache-2.0
language:
- ar
pipeline_tag: image-classification
datasets:
- arabic-handwritten-characters
metrics:
- accuracy
model-index:
- name: arabichar-v3
  results:
  - task:
      name: Image Classification
      type: image-classification
    dataset:
      name: arabic-handwritten-characters
      type: arabic-image-alphabets
      config: default
      split: train
      args: default
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.9764
---
# Arabic Handwritten Character Classification

This model is implemented using a custom CNN model, named arabichar, for the classification of handwritten arabic characters (Hijaiyah) on [arabic handwritten characters](https://www.kaggle.com/datasets/mloey1/ahcd1) dataset.
It achieves the following results on the evaluation set:
- Loss: 2.4150
- Accuracy: 0.9764