Team IT Security (TSEC) Nachrichtenportal Logo

💾 trunk/319ee4ea19c03438d7ae3c585bc65b11fb9dd266: Add PyTorch QuACK GEMM epilogue adapter (#186310)


Nachrichtenbereich: 💾 Downloads
🔗 Quelle: github.com

Pull Request resolved: #186310
Approved by: https://github.com/slayton58
ghstack dependencies: #186284 [Weiterlesen]

Sharing is caring on Social Media

Reddit
Telegram
LinkedIn
Xing
Twitter
Whatsapp
Facebook
Flipboard

💾 trunk/46931e0e655d691efc0bf431ab0fe920e7db871a: Remove plain asserts in torch/_dynamo/source.py (#182171)


📈 575.47 Punkte
💾 Downloads

🔧 DeepGEMM Essentials: High-Performance FP8 Matrix Multiplication


📈 453.41 Punkte
🔧 Programmierung

💾 trunk/873eaeb74249197dc7de29303aed4fcf47823949: [BE] Build all CUDA wheels per (arch, cuda) on a single runner (#184045)


📈 411.05 Punkte
💾 Downloads

💾 ciflow/trunk/182326: Update on "Add tearDown checks for torch function state leaks"


📈 411.05 Punkte
💾 Downloads

🔧 Writing High-Performance Kernels in TileLang, from GEMM to MLA


📈 404.68 Punkte
🔧 Programmierung

🔧 From DeepSeek to Quack: When the Dream of Distributed DuckDB Started to Feel Real


📈 382.26 Punkte
🔧 Programmierung

💾 trunk/33004619eb042c0a6b85db3e2e17a6409b89219d: [DTensor] Use explicit hints for unbacked sharding (#183545)


📈 369.95 Punkte
💾 Downloads

🔧 Shai-Hulud Malware in PyTorch Lightning: What Actually Happened and How to Check Your Environment


📈 315.14 Punkte
🔧 Programmierung

🔧 DeepSeek DeepGEMM 中文讲解


📈 299.99 Punkte
🔧 Programmierung

💾 ciflow/torchtitan/186020: Reintroduce ptxas bundling for CUDA 13.0 binaries


📈 287.74 Punkte
💾 Downloads

💾 trunk/419b2d0b867ae2c6481a0f87c1e99e02ea4b3dc0: Remove plain asserts in torch/_dynamo/compiled_autograd.py (#182164)


📈 287.74 Punkte
💾 Downloads

💾 trunk/55811c3e7b01509262d4e7862c15eb17b439ceae: [ATen][FakeTensor] Handle unbacked dims in folded matmul (#183397)


📈 274.03 Punkte
💾 Downloads

🔧 How I Achieved 2.9 PyTorch Training Speedup with One Line of Code


📈 260.33 Punkte
🔧 Programmierung

💾 trunk/826ab493be5714d267a00badb3773e0d9b78e00f: [Dynamo] Remove has_unpack_var_sequence (#185153)


📈 260.33 Punkte
💾 Downloads

🔧 Convert Tensorflow to Pytorch Model. in 2025?


📈 260.33 Punkte
🔧 Programmierung

💾 trunk/151e7e0cbb86bb29c72bc5b116670ed3221cb5e5: [inductor] Support InductorBenchmarker for non-CUDA device (#184567)


📈 246.63 Punkte
💾 Downloads

💾 trunk/12bca9da60b36ba38935e1305bd266bdaf1f5d1e: Run EC2 and OSDC in parallel on trunk via ciflow tag pushes (#183482)


📈 246.63 Punkte
💾 Downloads

🔧 DuckDB 1.5.3 Adds Quack Client-Server, SQLite Gets Cypher Graph Extension


📈 244.65 Punkte
🔧 Programmierung

🔧 AWS re:Invent 2025 - Break through AI performance and cost barriers with AWS Trainium (AIM201)


📈 232.93 Punkte
🔧 Programmierung

💾 viable/strict/1780616745: [inductor] Fix module exhaustion during exhaustive tuning (#184285)


📈 212.38 Punkte
💾 Downloads

🔧 TensorFlow vs PyTorch: The Real Difference Isn't Accuracy


📈 212.38 Punkte
🔧 Programmierung

💾 viable/strict/1780614985: Build torch._C extension module and version.py via CMake (#180243)


📈 205.53 Punkte
💾 Downloads

💾 viable/strict/1780178955: Use SVE `exp_u20`/`fexp_u20` for SVE128 (#184341)


📈 205.53 Punkte
💾 Downloads

💾 trunk/eaa777900b14e1aa6ffc1959cbcdb055d5de3e75: Update third_party/kineto submodule to 7739225 (#182973)


📈 205.53 Punkte
💾 Downloads

💾 trunk/9954ca57923e230e51901bf08265f69c47969702: Remove plain asserts in torch/_dynamo/ (more top level files) (#182162)


📈 205.53 Punkte
💾 Downloads

🔧 PyTorch vs TensorFlow 2025: Which one wins after 72 hours?


📈 198.68 Punkte
🔧 Programmierung

🔧 Terkuak! Pertarungan Sengit TensorFlow vs PyTorch: Siapa Sebenarnya Raja AI yang Wajib Anda Kuasai?


📈 198.68 Punkte
🔧 Programmierung

🔧 Apple Silicon's AI Ceiling Is Higher Than You Think


📈 188.44 Punkte
🔧 Programmierung

🔧 The Great AI Convergence: PyTorch vs. TensorFlow in 2026


📈 184.97 Punkte
🔧 Programmierung

💾 trunk/9c77ff51b6b2a30dc0964289e54c55b3f1bcdf13: Fix functional accumulate_grad state modeling (#184082)


📈 184.97 Punkte
💾 Downloads

💾 trunk/2a9d5cb7a5d6b8a870f4a992f570599c38e4ef68: [BE] Build all aarch64 + x86 CPU wheels on a single runner (#183931)


📈 178.12 Punkte
💾 Downloads

💾 trunk/01399c7c630cea29aeba8b64ed17be1caf549b4c: [ROCm][CI] add gfx950 CDNA2OrLater() in common_cuda.py (#183599)


📈 178.12 Punkte
💾 Downloads

🔧 Qlib - 模型训练和预测(Model Training & Prediction)


📈 178.12 Punkte
🔧 Programmierung

🔧 Project of the Week: PyTorch


📈 178.12 Punkte
🔧 Programmierung

🔧 How to Read GPU Profiling Logs: A Ground-Up Guide


📈 174.36 Punkte
🔧 Programmierung