Pinned Loading
-
-
-
Neural_Network_From_Scratch
Neural_Network_From_Scratch PublicFrom scratch NumPy implementation of a neural network library - dense layers, backprop, optimizers, no ML frameworks
Jupyter Notebook
-
AUTOML
AUTOML PublicCLI AutoML tool that takes any CSV + task type (classification/regression/clustering) and automatically preprocesses the data and benchmarks 25 hand-tuned scikit-learn/XGBoost models built to learn…
Python
-
optimizer-diagnostics
optimizer-diagnostics PublicA from-scratch NumPy neural network comparing SGD, Adagrad, RMSprop, and Adam on Fashion-MNIST — not just loss curves, but effective learning rate, gradient noise scale, and minima sharpness (via H…
Jupyter Notebook
-
bengio-neural-probabilistic-language-model-numpy
bengio-neural-probabilistic-language-model-numpy PublicBengio et al. (2003) neural probabilistic language model implemented from scratch in NumPy, with hand-derived backprop and a gradient check. No PyTorch.
Python
If the problem persists, check the GitHub status page or contact support.