ExecuTorch adds WebGPU backend: run PyTorch models in-browser, up to 10x faster than CPU
PyTorch · x · 2026-11-11
PyTorch announced the ExecuTorch WebGPU backend: instead of converting models to another framework's format, you export once with torch.export and the same .pte file runs in WebGPU browsers like Chrome and Safari — fully on-device, so prompts, photos, and audio never leave the tab.
- Verified models: Llama 3.2, Qwen3, Gemma 3, SmolVLM, SAM 2, Depth Anything V2, D-FINE, Real-ESRGAN, and Whisper.
- Supports on-device LoRA and DiReFT fine-tuning.
- Benchmarks on Apple GPU + Chrome show 4-7.4x prefill and 1.8-10x decode speedups over the CPU (XNNPACK) backend for Llama 3.2 1B and Qwen3-0.6B across context depths.
- A WebGPU Playground is available to try the models.
More from Infra
- SpaceX locks 100+ GW of solar cell equipment from SINGULUS for Starmind AI satellites — ns123abc · 2026-11-11
- CoreWeave upgrades to CNCF Platinum member, runs its GPU cloud entirely on Kubernetes — altryne · 2026-11-11
- SpaceX teams with Qualcomm to build AI-native base stations, eyeing first AI carrier — elonmusk · 2026-11-11
- Prism capability extraction: specialists carved from Qwen 27B at <20% memory match full model — tokenbender · 2026-11-11
- Apple reportedly building Harbor app to turn Macs into private AI inference servers with Thunderbolt 5 clustering — BenBajarin · 2026-11-11
- Agent Arena: Gemini 4 Carbon Gets More Cost-Competitive as Tasks Get Harder — rseroter · 2026-11-11