PyTorch

Track PyTorch development. The machine learning framework powering AI research.

About this developer podcast

PyTorch turns GitHub activity into short audio updates.

Episodes cover commits, pull requests, issues, and release work.

Use this page to listen, subscribe, and find related developer podcasts.

This show page is the public hub for the podcast. It has one stable listen URL.

Use it to find RSS feeds, recent episodes, related shows, and older episode links.

If the project is quiet, new episodes may pause until there is useful activity to summarize.

Recent episodes from PyTorch

  1. Distributed Reliability and Compile Overhead Cuts

    Distributed fixes dominate, tackling cross-group signal interference, store hangs at scale, and idle communicator memory. A parallel Dynamo and Inductor push trims warm-cache compile overhead while fixing silent pattern-rewrite drops.

  2. Compiler Memory Cleanup and Distributed Savings

    A large Inductor stack tackles memory retention and recompilation costs while fusing more work into matrix-multiply kernels, and distributed changes cut FSDP2 complexity and collective test time. The pattern is efficiency: less memory…

  3. Dynamo Object Defaults, Release Cleanup, and Test Visibility

    Dynamo dominates with Python 3.10 removal and true object defaults for repr, string, and comparison to cut graph breaks. Release 2.15 reverts, Inductor strict numerics fixes, and new crash reporting round out a reliability-focused day.

  4. Inductor Correctness Fixes and Regional Graph Control

    Inductor fixed silent numerical divergences in pattern matching and strict-mode operations while adding safer performance tuning. Dynamo consolidated its internal tracking logic, and distributed tests got major speedups as the 2.15…

  5. All-Gather Showdown and Stack-Aware Merges

    Fully sharded training is weighing three competing designs for its next all-gather contract, while merge tooling gains full support for GitHub-native stacks. Apple-silicon speedups and empty-input crash fixes round out a…

  6. Weekly Recap - FSDP Contracts, MPS Speedups, and Compiler Fidelity

    FSDP2 all-gather redesign debate led 50 pull request activities and 30 commits, alongside major MPS speedups and linalg fixes. The weekly pattern was correctness at the edges across distributed training, Apple silicon, and compiled paths.

  7. Compiler Parity Push

    Dynamo, Inductor and ahead-of-time packaging all moved to match eager Python behavior. The work tightens numeric fidelity, reduces graph breaks, and fixes threading and memory ownership.

  8. Distributed Hang Fixes and Compiler Hardening

    Distributed collectives dominated activity with deadlock and silent-hang fixes bringing N-C-C-L-2 toward parity. Dynamo, Export and Inductor also saw a cluster of correctness fixes, alongside a shift of agent triage to a read-only pi…

  9. Low-Precision Grouped GEMM Push and Compiler Fixes
  10. Block Sharding for MoE, Collective Transparency, and Inductor Fixes
  11. GEMM Fusion and Portable CUDA Graphs
  12. Precompile Groundwork and Compiler Reliability Fixes
  13. Compiler Correctness and Precompile Hardening
  14. Weekly Recap - Precompile Scale, Mac Reliability, and Compiler Fixes
  15. Precompile for Serving, Faster First Compile, MPS Correctness
  16. CI Pod Cleanup and Compiler Correctness
  17. Dynamo Fidelity, MPS Scale, and CI Moves
  18. SVD Fixes, Compiler Safety, and Export Hardening
  19. Device-Agnostic Tests, Inductor Reliability, and GPU Partitioning
  20. Precompile Goes Default, Dynamo Tightens Correctness
  21. Dynamo Guards Hardening and Precompile Push
  22. Weekly Recap - Precompile Goes Public and Guards Become Portable
  23. Precompile Interface and Guard Packaging Hardening
  24. The Great Distributed Test Speedup
  25. Compile Regions Get Finer Control
  26. Test Infrastructure Gets a Deep Clean
  27. Sharding, Fake Tensors, and the Great Test Cleanup
  28. Correctness Fixes and the AOT Compile Hardening Push
  29. Compile-On-One-Rank Gets Serious About Device Portability
  30. Weekly Recap - Compiler Correctness and MPS Backend Consolidation
  31. Apple Silicon Hardening and Dynamo's Global State Cleanup
  32. Hardening the Compile-and-Ship Pipeline
  33. Distributed Training Takes Control of Its Own Timing
  34. Guard Serialization Gets a Precision Pass
  35. Inductor's Reduction Fusion Overhaul
  36. Hardware Independence Takes Center Stage
  37. Hardware-Agnostic Cleanup and MPS Correctness Push
  38. Weekly Recap - Multi-Rank Compilation and Backend Correctness
  39. One Rank to Compile Them All
  40. Dynamo Closes Its Torch Function Gaps
  41. Precompile Gets a Real Foundation
  42. Cleaning Up Exceptions, Symbolizers, and Inductor's Blackwell Push
  43. Distributed Transports, Compiler Correctness, and the Aches of Reference Counting
  44. Memory Safety and Lifetime Fixes Across CUDA and Distributed
  45. Precompile Lands, and the Correctness Sweep Continues
  46. Weekly Recap - Precompile Lands, and Correctness Sweeps Across MPS and CUDA
  47. Batching Rules Get Unified, Precompile Gets Sharper
  48. Undefined Contracts and the MPS Coverage Push
  49. A New CuteDSL Reduction Engine and the Great Test Untangling
  50. ROCm Ten Point Oh and the Precompile Stack
  51. Hardening the Merge Pipeline and the Attention Stack
  52. FlyDSL Lands and the GEMM Compiler Stack Grows
  53. AOT Autograd Tightens Its Grip on Backward Tangents
  54. Weekly Recap - Compiler Correctness and the Long Tail of Cleanup
  55. Precompile Guards Get a Rewrite
  56. Precompile Hardening and the Great Cleanup
  57. Precompile's Guard Overhaul
  58. Compiler Stacks, Symmetric Kernels, and Build Hygiene
  59. Precompile Grows Up, and the Great Typo Sweep
  60. Correctness Cleanup and a Suspicious Cluster
  61. Numerical Correctness Sweep
  62. Weekly Recap - Release Stabilization and Dynamo's Object Model Overhaul
  63. Dynamo's Great Untangling
  64. Attention Backend Correctness and CI Coverage Gaps
  65. The Great Hardware Decoupling
  66. The Great Device-Agnostic Test Rewrite
  67. Release Branch Cleanup and the Great Test Decoupling
  68. The Great Test Reclassification
  69. The Great Test Suite Decoupling
  70. Weekly Recap - Test Infrastructure Overhaul and Correctness Fixes
  71. The Great NaN and Negative Zero Cleanup
  72. Closing the GPU Test Coverage Gap
  73. The Great Device-Agnostic Test Migration
  74. Hardening Distributed Communication and Cutting Legacy Support
  75. Test Debt, GPU Correctness, and the Long Tail of Cleanup
  76. Distributed Backend Correctness Overhaul
  77. Boundary Checks and the Hardware-Agnostic Test Push
  78. Weekly Recap - Silent Correctness Fixes Take Center Stage
  79. Silent Corruption Sweep
  80. Compiler Internals Get Hardened Ahead of Bigger Changes
  81. Distributed Backends Get Serious About Parity
  82. A GEMM Epilogue Megastack and Distributed Hardening
  83. Cleaning Up the Compiler's Noise
  84. Crash Fixes and NCCL's Next Chapter
  85. Compiler Correctness Meets Test Portability
  86. Weekly Recap - Performance Fixes and Toolchain Modernization
  87. Toolchain Upgrades and Correctness Fixes Across the Stack
  88. Native Kernel Push and Distributed Backend Parity
  89. Correctness Hardening Across Inductor and AOTI
  90. The NVGEMM Epilogue Fusion Marathon
  91. Distributed Correctness Push
  92. CUDA Graphs Get a Lifecycle, Distributed Gets Cleaned Up
  93. One Engineer, Many Kernels
  94. Weekly Recap - Cleanup, Correctness, and a Random Number Overhaul
  95. Cleanup Sweeps and a New RNG Foundation
  96. Hardening Memory and Lifetime Management Across GPU Backends
  97. Correctness Fixes and Reverts Take Center Stage
  98. Distributed Backends and Overflow Fixes Take Center Stage
  99. Packaging Fixes and Bounds-Check Sweep
  100. Autograd Overhead Cuts and a New GEMM Backend Migration

Episode archive for PyTorch

  1. Hardening Boundary Checks and MPS Correctness
  2. Weekly Recap - Metal Kernels, Compile Reliability, and Dynamo's Long Tail
  3. Apple Silicon Gets a Metal Overhaul
  4. Dynamo's Great Type-System Cleanup
  5. Dynamo's CPython Test Push and a Rocky Exception-Handling Rewrite
  6. Distributed Backend Overhaul and a Wave of Cleanup
  7. The Great Runner Migration, and a Distributed Rollback
  8. Distributed Cleanup and CI Infrastructure Overhaul
  9. Correctness Fixes Sweep the Compiler Stack
  10. Weekly Recap - Release Hardening and Correctness Cleanup
  11. Determinism and Correctness Sweep
  12. Silent Correctness Bugs Take Center Stage
  13. FlexAttention's Rough Week
  14. Accelerator Backends and Memory Management
  15. Weekly Recap - Release Stabilization & Core Improvements
  16. GEMM Optimization and Core Stability Fixes
  17. C++20 Migration and Infrastructure Improvements
  18. Release Infrastructure and Compiler Stability Fixes
  19. Header Reorganization and Dynamo Improvements
  20. Infrastructure Organization and Core API Cleanup
  21. Dynamo Correctness and Build Infrastructure
  22. Performance Optimizations and Backend Improvements
  23. Weekly Recap - Compiler Optimization & Reliability
  24. Dynamo Compilation Stability and Performance
  25. Dynamo Stability and Inductor Optimizations
  26. Compiler Performance and Distributed Computing Infrastructure
  27. Distributed Naming Overhaul and Compiler Fixes
  28. Dynamo Stability and Compilation Robustness
  29. Compiler Numerical Accuracy and Index Expression Fixes
  30. Compiler Error Handling and Stability Fixes
  31. Weekly Recap - Stability and Error Handling Improvements
  32. Weekly Recap - Code Quality & Infrastructure Improvements
  33. Weekly Recap - Dynamo Enhancements & Critical Fixes
  34. Optimization Engine Revamp and Hash System Breakthrough
  35. Operator Improvements and Build System Updates
  36. Weekly Recap - Reverts and Infrastructure Updates
  37. Inductor Improvements and Bug Fixes
  38. Optimization Improvements and Cache Unification
  39. Weekly Recap - Performance Optimizations and Code Cleanup
  40. Weekly Recap - Infrastructure Stability & Mixed Precision
  41. Dynamo Improvements and Distributed Computing Fixes
  42. Weekly Recap - Performance Optimization and Infrastructure
  43. The Debugging Detective Story
  44. The Great Rollback and Recovery
  45. Exception Handling Revolution & Stateless RNG Arrives
  46. Cross-Platform Expansion and Developer Experience Wins
  47. Building Stronger Foundations
  48. Spring Cleaning & Major Infrastructure Upgrades
  49. AOT AutoGrad Fixes and Cross-Platform Polish
  50. Building Bridges - Distributed Computing Gets a Major Upgrade
  51. Profiling Power-Ups and Infrastructure Smoothing
  52. Fixes, Reverts, and Moving Forward
  53. The Infrastructure Acceleration Edition
  54. Lanczos Interpolation Breakthrough
  55. Stream Management Mastery & RNG Fixes
  56. Matrix Math Gets a Speed Boost
  57. Under the Hood Improvements and Future-Proofing
  58. Complex Math Gets Smarter & Build Improvements
  59. Memory Optimization Revolution
  60. Testing Gets Smarter and Graphs Go Universal
  61. Polish & Performance Day
  62. Distributed Computing Gets Real - Compilation, Clustering, and Convolutions
  63. Performance Revolution and Developer Experience Upgrades
  64. Windows Testing Gets Flexible & Dynamic Shapes Take Flight
  65. Metal Shaders Get a Precision Fix
  66. The Testing & Error Handling Polish Episode
  67. Stream Safety and Performance Wins
  68. Subclass Evolution and Memory Management Improvements
  69. Performance Tuning and Code Health Day
  70. Variable-Length Attention Gets Supercharged
  71. Spring Cleaning and Performance Boosts
  72. Stream Wizardry and Symbolic Shapes Magic
  73. CI Optimizations and Cross-Platform Fixes
  74. Spring Cleaning and Precision Fixes
  75. Memory Safety Fixes and Development Velocity
  76. Speed Wins and Better Error Messages
  77. Distributed Computing Gets Smarter
  78. Distributed Computing Gets Smarter & Vision Models Get Lightning Fast
  79. Release Dance and Rapid Recovery
  80. Performance Wins and Stability Fixes
  81. Distributed Computing Gets Smarter
  82. Valentine's Day Cleanup and Distributed Computing Love
  83. The Day of Rollbacks and Second Chances
  84. TPU Integration and the Dance of Reverts
  85. The Performance Optimization Sprint
  86. The Great Performance Revolution - Tests Run 70% Faster!
  87. Bug Fixes and Performance Wins
  88. The Great Test Speed Revolution
  89. Cleanup and Optimization Day
  90. Testing Cleanup and Pattern Matching Progress
  91. Type Safety Revolution and Infrastructure Cleanup
  92. Backend Flexibility Revolution
  93. The Great Configuration Cleanup & XPU Expansion
  94. Hardware Expansion and Developer Experience Polish
  95. Backend Harmony and Memory Magic
  96. Spring Cleaning and Building Blocks
  97. Bytecode Magic and Buffer Management Mastery
  98. Kernel Optimization and Clean Code Victory
  99. FMA Optimization Focus and Debugging Improvements
  100. Developer Tooling Revolution
  101. Deep Dive into PyTorch's Core - Opaque Objects and Performance Wins

More public developer podcasts

Browse other public Podlog show hubs with crawler-readable RSS and episode links.

  1. Maestro Daily 91 episodes
  2. Videorc Daily 84 episodes
  3. Navidrome Daily 114 episodes
  4. Next.js Daily 256 episodes
  5. Pi Mono 136 episodes
  6. OpenClaw 213 episodes
  7. RuView 168 episodes
  8. Godot Daily 201 episodes
  9. Agora Next Updates 271 episodes
  10. Rails Daily 255 episodes
  11. Home Assistant Daily 260 episodes
  12. Linux Kernel Daily 234 episodes
  13. React Daily 151 episodes
  14. Jabref Daily 102 episodes
  15. BlocksBeyondTheStars Daily 82 episodes
  16. Jabref Daily 103 episodes
  17. Next Js Daily 91 episodes
  18. Headroom Daily 100 episodes
  19. Agent Of Empires Daily 91 episodes
  20. Frigate NVR Updates 238 episodes
  21. Linux Kernel 228 episodes
  22. Homebrew 251 episodes
  23. TypeScript 140 episodes
  24. NativeScript iOS Daily 42 episodes
  25. Ruby on Rails 244 episodes
  26. Redis 219 episodes
  27. Next.js 253 episodes
  28. Vue.js 191 episodes
  29. Rust 223 episodes
  30. Ruby Core Updates 246 episodes
  31. Python 236 episodes
  32. Go 236 episodes
  33. VS Code 254 episodes
  34. TailwindCSS 147 episodes
  35. Node.js 248 episodes
  36. Django 240 episodes
  37. React Native 197 episodes
  38. PostgreSQL 239 episodes
  39. Kubernetes 249 episodes
  40. LangChain 234 episodes
  41. Ollama 238 episodes
  42. Openharness Daily 63 episodes
  43. Shannon 104 episodes
  44. Rimworld ChickenBatteryCage Daily 1 episode
  45. Eleviewer Daily 1 episode
  46. iCloud Photos Downloader 36 episodes
  47. Buzz Transcription 99 episodes
  48. The Algorithm Daily 20 episodes
  49. AtomVM Daily 20 episodes
  50. Agora Next Daily 16 episodes
  51. Onlook Design Updates 34 episodes
  52. Energy Inc Ace Mate Learning Daily 2 episodes
  53. tiny-gpu Daily 20 episodes
  54. OpenAI Skills 52 episodes
  55. Crush Daily 1 episode
  56. ControlNet Daily 0 episodes
  57. FinceptTerminal Daily 0 episodes