Spaces:
Running
Running
Commit History
6 days ago
-
Update index.html
Verified signature8b7adc3 -
Update js/agate.js
Verified signature7f7fbbf -
Update js/app.js
Verified signature498391f -
Upload models/vae_encoder/model_sd15.onnx
Verified signatured2931d4 -
Upload models/vae_encoder/model.onnx
Verified signature65fc8c5 -
Update js/app.js
Verified signature88143fe -
Update index.html
Verified signature66e023b
18 days ago
-
Why Agate: general image generation as a benchmark-friendly proxy
Stefatorus +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018EtDckYfNgMbkwqVmTsdhadf091f0 -
Why Agate: scope (glyph/symbol and vector-native font generation), released to share our architectural efforts
Stefatorus +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018EtDckYfNgMbkwqVmTsdha8d1a72e -
Faster WebGPU sampling: fp16 compute, GPU-resident steps, non-blocking thinker preview
Stefatorus +1: Claude Opus 5.5- 002/003 and 001 load the v2 build from webgpu/v2/ in each model repo: one graph run per Euler step (CFG mix + update inside the graph), z stays on the GPU between steps, fp16 compute (fp32 for graph I/O, CFG/Euler, timestep/resolution/count embeddings and Range), 1x1 conv + depth-to-space instead of ConvTranspose. - 003 @ 512: 275 -> 161 ms/step, 14 -> 8.1 s/image; 256 px and 001/002: ~6 -> ~3.4 s (RTX 4060, Edge). - Show thinker: async readback, colouring in a Web Worker, frames dropped while busy; no cost to sampling. - "Fast" toggle: 25 steps (~4.1 s at 512); default stays 50. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018EtDckYfNgMbkwqVmTsdha146ba97 -
Preview 002 and 003 (512 px), version selector, Show thinker, AI-output marking
Stefatorus +1: Claude Opus 5.5- Model selector (003 default for a first visit, last choice remembered per browser); 003 at 512 or 256 px with its prompt pipeline (normaliser, spelled-out quoted text, count code, "no X" -> negative) and SD3 shift. - The ONNX files now load from each release's model repo (Logolabs/agate-preview-00X, webgpu/); the graphs also output the thinker plan. models/ (the original 001 build) stays for ?models=./models/. - Show thinker (default on, ~2-4% per step): the 16 x 16 plan (PCA colours, as the ComfyUI live preview) and the expected image x1 at every step. - Every image is marked: invisible-watermark dwtDctSvd (payload AGATE + release, bit-identical JS port), PNG provenance text chunks on save (no prompt), and a visible "AI-generated" note. - Parity fixtures for 002 and 003, and a 38-prompt golden test of the prompt pipeline (?golden=1). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018EtDckYfNgMbkwqVmTsdha2dc8d66
21 days ago
-
Count every model load, like from_pretrained
Verified signatureaaa2328 -
Read the model config once per browser for the download counter
Verified signatureb0dc8c9 -
Count first-time model loads on the Logolabs/agate-preview-001 download counter
Verified signature1c22c34
22 days ago
-
Declare the model this Space uses
Verified signature007169e -
Link the now-public model card
Verified signaturef3b1614 -
README: speed line
Verified signature936aea5 -
Remove template style.css
Verified signatureebfded8 -
Agate WebGPU: static in-browser demo with ONNX models
Verified signature7a7d8dc -
initial commit
Verified signaturec55a31d