numpy

"I started with Python, learned some Pandas and NumPy, worked with a dataset from Kaggle, and tried Matplotlib (still pretty bad at it 😅)."

"In NumPy, four arithmetic operations between arrays can be done intuitively with +, -, *, /. ... Additionally, by using sum() on an array, you can find the sum of all elements within that array, and for 2D arrays and higher, you can specify an axis to perform vertical or horizontal additions."

"I decided to challenge myself by setting up the environment using uv, rather than Anaconda or venv, because its incredible speed has earned it the reputation of being the next-generation Python package manager."

"My goal is to understand and implement ML algorithms, preprocess data, and use libraries like NumPy, Pandas, and scikit-learn."

"PilotNet で自動運転カートを End-to-End で走らせてみた"

"A determinant tells us two useful things: • Whether a matrix can be inverted • How a matrix transformation changes area"

"I also explain the matrix inverse, which is similar to division with numbers."

"Tensor Spy is a free webapp that lets you quickly inspect the contents of numpy & pytorch tensors locally (your tensors are not uploaded to any servers)."

"I’m still learning, but one thing I understood quickly is that machine learning is not just about coding models. A big part of it is cleaning data, analyzing patterns, and understanding the problem you’re trying to solve."

"Ultimately, when the scratch implementation of XOR was completed, I was filled with a sense of accomplishment, like defeating the origin of the AI winter that had lasted for a long time, and a sense of accomplishment that I had implemented in code in 2026 the thoughts that Rosenblatt had tried to achieve in the 1960s but could not realize, and the background that I had once been frustrated by XOR also overlapped, and I felt very deeply moved."

"The most reliable and effective approach to this problem is to restore the integrity of the environment and recompile dependent packages in accordance with the current NumPy."

"Instead of absolute values (price itself), "percentage change (rate of change)" or "logarithmic return" is used for input and output, making it a more versatile model."

"By correctly combining NumPy's reshape and transpose, PixelShuffle can be fully replicated in a pure NumPy environment without relying on external libraries."

"This article implements a Deep Neural Network (DNN) using only NumPy, without frameworks (PyTorch or TensorFlow), to create a model that predicts the Nikkei stock average from the exchange rate (USD/JPY)."

"I found this free guide that walks through building a simple deep learning library from scratch using just NumPy."

"The program in this article is running on Python version 3.13. Also, the version of numpy is 2.3.5."

"MNIST digit recognition is done using Python without using frameworks, etc."

"This article highlights the development of SmallPebble, a minimalist deep learning library written from scratch in NumPy."

"The article's program is running on Python version 3.13 and numpy version 2.3.5."

"This series dissects the inner workings of LLMs, from full scratch implementations with Python and NumPy, to cutting-edge techniques used in Qwen-32B class models."

"When handling multidimensional arrays of 3 or more dimensions, imagine a 'solid' in your head..."

"The article aims to help users not get confused with NumPy's multi-dimensional array operations."

"The article's context provides no direct quotes."

"The article likely explains how to build a CNN using only NumPy."