aditya30394 / handwritten-digit-recognition-mnist

Classifying gray scale images of handwritten digits (28 × 28 pixels) into their 10 categories (0 through 9) using deep learning

Handwritten-Digit-Recognition-MNIST

Classifying gray scale images of handwritten digits (28 × 28 pixels) into their 10 categories (0 through 9) using deep learning

Results/Logs

Train dataset size: 60000

Test dataset size: 10000

Train Epoch: 0 Loss: 0.413797 Accuracy: 51804/60000 (86.34%)

Train Epoch: 1 Loss: 0.224540 Accuracy: 55131/60000 (91.89%)

Train Epoch: 2 Loss: 0.487874 Accuracy: 56058/60000 (93.43%)

Train Epoch: 3 Loss: 0.099301 Accuracy: 56761/60000 (94.60%)

Train Epoch: 4 Loss: 0.067272 Accuracy: 57246/60000 (95.41%)

Train Epoch: 5 Loss: 0.191264 Accuracy: 57624/60000 (96.04%)

Train Epoch: 6 Loss: 0.143393 Accuracy: 57959/60000 (96.60%)

Train Epoch: 7 Loss: 0.059414 Accuracy: 58192/60000 (96.99%)

Train Epoch: 8 Loss: 0.362808 Accuracy: 58372/60000 (97.29%)

Train Epoch: 9 Loss: 0.043404 Accuracy: 58571/60000 (97.62%)

Test set Results: Average loss: 0.0007, Accuracy: 9714/10000 (97.14%)