Unlocking AI at Scale for PyTorch Developers

on YouTube ↗08m

" summary

Qualcomm outlines a full edge stack: hardware (CPU GPU NPU), Qualcomm AIHub for web-enabled model testing and deployment, and ExecutorTorch reaching GA on the Qualcomm Hexagon NPU. Claims include initial support for 25+ generative models, live demos that accept custom prompts and images, measured improvements of roughly 2–4x token rate versus CPU and up to 75% faster model load, and tooling that converts PyTorch → ONNX → Qualcomm runtimes for scaled edge deployments.

AI-written summary · about

" related talks

• 29mQualcomm® AI Stack for developers and extension to on-device AIAmber Branwell, Jackie Donoghue, Fred Balsiger, Vinesh Sukumar · Microsoft Build 2023
• 37mHow to Optimize, Validate and Deploy ML Models On DeviceAshvin Kumar, Meghan Stronach · droidcon Berlin 2025
• 14mAccelerating On-Device AI DevelopmentSiddhika Nevrekar · droidcon New York 2024
• 26mHow to use Qualcomm AI HubShreya Jain, Meghan Stronach · droidcon New York 2024
• 15mBuilding Gen-AI experiences using LLM and LVMAnkit Kapoor · Microsoft Build 2024
• 22mIntroducing ExecuTorch from PyTorch Edge: On-Device AI...Mergen Nachin, Orion Reblitz-Richardson · PyTorch Conference 2023
• 16mLightning Talk: Efficient Inference at the Edge: Performance You Need at the Lowest...Felix Baum · PyTorch Conference 2023
• 54mHarnessing On-Device AI to Transform the Smartphone ExperienceVidyasagar Rao · TWIMLcon: AI Platforms 2021