Short description for quick search
-
Updated
Jan 31, 2019 - Python
Short description for quick search
Finite difference methods.
Implementation of all basic algorithms needed in Deep Learning
Curso Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization. Segundo curso del programa especializado Deep Learning. Este repositorio contiene todos los ejercicios resueltos. https://www.coursera.org/learn/neural-networks-deep-learning
Building Convolution Neural Networks from Scratch
Neural_Network for handwritten numbers detection in Octave (free Matlab)
FloydHub porting of deeplearning.ai course assignments
Build your deep learning algorithms with confidence
Word2Vec (Skip-gram with negative sampling) implemented from scratch in NumPy: derived gradients verified numerically, subsampling, and noise sampling. Matches gensim on WordSim-353 (0.364 vs 0.365).
A transformer language model in one file of numpy, backward pass written by hand and checked. Train it on one paragraph, or on your own text.
零依赖可验证 AI 内核:tape 自动微分 + RoPE/GQA/SwiGLU/MoE + KAN,有限差分梯度检验全绿。作者 晨星(CJX0712)
Conv Forge · 卷积神经网络锻造炉 — 单文件零依赖,手写 2D 卷积(im2col+GEMM)/最大池化/反向传播,数值梯度检验对拍到机器精度,端到端训练真实收敛,10 项不变量全绿。作者:晨星 (CJX0712)
To associate your repository with the gradient-check topic, visit your repo's landing page and select "manage topics."