595 words
Building a Deep Learning Framework from Scratch in Python
training framework in pure Python and NumPy. No PyTorch, no TensorFlow, no autograd library. One dependency: NumPy for
Install
Latest in PyTorch, TensorFlow & JAX
188 words
CUDA-Oxide 0.2 Brings Early Improvements To Pure Rust CUDA Kernels
Phoronix - Categories Computers Display Drivers...Software Storage Operating Systems Peripherals CUDA-Oxide 0.2 Brings Early Improvements To Pure Rust CUDA Kernels Michael Larabel is the principal author of
75 words
Auto-Generated CUDA Kernels Need Kernel-Level Validation
An LLM-written kernel benchmarked 38% faster on a microbench. Here is what kernel-level validation... Tagged with ai, machinelearning, gpu, performance.
1,606 words
Keras vs TensorFlow: Key Differences & Which to Choose (2026)
learning faces the same confusion: should I learn Keras or TensorFlow? The question itself is misleading because Keras and TensorFlow are not two separate choices competing for the same job. Keras is a high-level API specification written entirely in Python.
1,614 words
PyTorch vs TensorFlow: Key Differences and Benchmarks
project in 2026, you have likely asked: should I use PyTorch or TensorFlow? The short answer: PyTorch is the default for most new projects. TensorFlow remains the stronger choice for mobile deployment,...you are deploying it. Over 25121 companies use TensorFlow in…
225 words
Hugging Face Launches PyTorch Profiler Tutorial Series for Performance Optimization
published the first installment of 'Profiling in PyTorch,' a beginner-friendly series on reading profiler...execution chain—from Python function calls down to CUDA kernels. The series uses NVIDIA A100-SXM4-80GB GPUs...modern GPU acceleration patterns. torch.profiler is PyTorch's built-in performance…
623 words
I Built a Tiny Neural Network From Scratch in Python — No PyTorch
So I decided to build one myself. No PyTorch. No TensorFlow. No Keras. Just Python + NumPy + the mathematics behind a...wasn't to build a production-ready deep learning framework. The goal was to understand what actually...rate. After building this manually, frameworks…
609 words
Calling CUDA from Go without cgo
high-throughput data work now runs on NVIDIA GPUs through CUDA. The problem is that most CUDA access from application code still lives in the Python world. This post is about why calling CUDA from Go matters, why cgo is often…
264 words
PyTorch vs TensorFlow: Which Should Beginners Learn First?
Introduction If you’re starting...journey into Deep Learning, you’ll eventually face a common question: Should I learn PyTorch or …
301 words
KerasFormers: Keras 3 collection of pretrained models
KerasFormers is an open-source library of pretrained models built in Keras 3, with one codebase designed to run on JAX, PyTorch, or TensorFlow. Its Product Hunt listing presents that portability...inference service. The official overview describes pure-Keras ports whose…
PyTorch, TensorFlow & JAX
Deep-learning frameworks and releases, CUDA programming and kernels, Triton, distributed training and performance tuning.
- 92 Tracked terms
- Last 6 months Feed window
What this topic collects on
An article joins this feed when it matches these terms. Each one is also a search of its own.
- +52 more