[KongchangAI]
Benchmark

Arabic Digits MNIST

An Arabic handwritten digit dataset commonly used to test and extend MNIST-related models.

Timeline (last 90 days)

Nov 2

The third model uses the Arabic Digits MNIST dataset, with an architecture of 784→20→20→15→10, cross-entropy loss plus plain gradient descent, an accuracy of about 96% after 2000 epochs, and uses JAX automatic differentiation to compute gradients.

Unverified50%

All Facts (1)

Unverified

The third model uses the Arabic Digits MNIST dataset, with an architecture of 784→20→20→15→10, cross-entropy loss plus plain gradient descent, an accuracy of about 96% after 2000 epochs, and uses JAX automatic differentiation to compute gradients.

50%

Source Articles

From Matrix Algebra to JAX: Building Neural Networks by Hand to Understand the Essence of Deep LearningNov 2