Demos¶
jamgrad includes five runnable demos that exercise core tensor ops, backpropagation, and neural-network training.
| Script | What it shows | Run |
|---|---|---|
demos/demo.py |
Core tensor ops, gradients, and graph export | python demos/demo.py |
demos/demo_chain_rule.py |
Chain-rule walkthrough with assertions | python demos/demo_chain_rule.py |
demos/demo_eigenvalues.py |
Power iteration with Tensor matmul |
python demos/demo_eigenvalues.py --matrix symmetric3 --iters 200 |
demos/demo_nn.py |
XOR training with a small MLP | python demos/demo_nn.py --epochs 1000 --lr 0.5 |
demos/demo_mnist.py |
MNIST classifier training | python demos/demo_mnist.py --dataset-size 5000 --epochs 20 |
Use the pages in this section for full options and output details per demo.