Topic

PyTorch

All digests tagged PyTorch

Suggested route

Topic journey

Move from earlier context through the available material to the newest signals. This route is a reading aid, not a difficulty ranking.

Start here

Build A Reasoning Model Scratch 2: Loading a Base Model, Text Generation, and KV Caching Sep 6, 2026 · 1:36:42 This deep dive covers loading pre-trained LLMs (Qwen3), implementing autoregressive text generation in PyTorch, and optimizing performance using KV caching and torch.compile for maximum inference speed.

Recent signals

Building a GPU Compiler from Scratch: Lowering PyTorch Graphs to CUDA via Multi-Level IRs Oct 8, 2026 · 53:59 Learn how ML compilers progressively lower PyTorch graphs to efficient CUDA kernels. We explore the IR stack: from PyTorch to Loop IR, Tiling, and Kernel IR, detailing fusion and scheduling. Why Reflection AI Is Making the Open Intelligence Bet - and What It Means for Deploying Models Oct 7, 2026 · 45:36 The shift to open intelligence means owning the entire AI stack. Customization now involves RL, environments, and controlling data/safety, moving beyond simple API calls.

GPU MODE

Building a GPU Compiler from Scratch: Lowering PyTorch Graphs to CUDA via Multi-Level IRs

Learn how ML compilers progressively lower PyTorch graphs to efficient CUDA kernels. We explore the IR stack: from PyTorch to Loop IR, Tiling, and Kernel IR, detailing fusion and scheduling.

Read article

Weights & Biases

Why Reflection AI Is Making the Open Intelligence Bet - and What It Means for Deploying Models

The shift to open intelligence means owning the entire AI stack. Customization now involves RL, environments, and controlling data/safety, moving beyond simple API calls.

Read article

Sebastian Raschka

Build A Reasoning Model Scratch 2: Loading a Base Model, Text Generation, and KV Caching

This deep dive covers loading pre-trained LLMs (Qwen3), implementing autoregressive text generation in PyTorch, and optimizing performance using KV caching and torch.compile for maximum inference speed.

Read article