Search the archive.
Find published work and reporting by title, description, article text or tag.
Refine results2
Also searching for Architecture, because “bim” is another name for it.
5 results for “bim”.
- Showcase
LFM2-MoE: An 8.3B Parameter Language Model Running Entirely in Your Browser via WebGPU
Liquid AI's LFM2-MoE loads 8.3B parameters (1.5B active) and generates text client-side through WebGPU. No server, no API key. A hybrid convolution-attention architecture running Mixture-of-Experts inference in a browser tab, powered by Transformers.js.
- Showcase
WebRF: The Longley-Rice Radio Propagation Model, Running in WebGPU Compute Shaders
Roman Liutikov ported the Longley-Rice ITM algorithm, a real-world radio propagation model used in cell network planning, entirely into WebGPU compute shaders. 1,200 lines of WGSL, a two-pass radial sweep architecture, and coverage maps that actually respect terrain.
- Showcase
Deferred Rendering in WebGPU: Sponza with 400+ Compute-Driven Point Lights
Georgi Nikolov's WebGPU renderer puts 400+ point lights in the classic Sponza atrium. Deferred shading, cascaded shadows, Hi-Z reflections, TAA. Clean architecture that handles WebGPU's explicit pipelines without drama.
- News
Google Brings LiteRT Inference to the Browser With WebGPU
Google has brought its LiteRT on-device inference stack to the browser, with XNNPACK on CPU, ML Drift over WebGPU, and experimental WebNN support for NPUs.
- Showcase
NullGraph: A Data-Oriented WebGPU Rendering Framework
Vikas Kumar Singh's NullGraph drops the scene graph and streams flat ArrayBuffers straight to the GPU. Strange attractors, Hopf fibrations, curl-noise fluids, and a test engine that lets you swap memory layouts mid-flight to watch the frame times move.
