Remove orphaned old-layout files (_shared vendor dup + old diagnostics)

Unified feature-complete build: clause-streamed voice + 260-char bubbles + PWA/offline SW + self-contained vendor

Q: canonical standalone chat rebuilt from apps/q engine (warm-KV persona-prefix reuse)

keep-warm pulse: hold GPU boost clock while typing so casual chat decodes at boosted rate

warmup uses decode() to precompile batched pipelines (cut first-msg TTFT)

fast first turn: warmup boosts clock + primes system-prompt KV (12s TTFT to ~0.5s) + static grounded greeting

discrete-GPU validation ?bench=discrete: bandwidth + live + spec-flip in one page

decode(): report live-path GPU ms/tok for clean CPU/GPU split

decode(): report live-path GPU ms/tok for clean CPU/GPU split

per-pass GPU trace ?bench=trace: name the non-weight overhead

per-pass GPU trace ?bench=trace: name the non-weight overhead

spec bench: warmup + short prompt + 192-tok decode-dominated (fix prefill confound)

live decode profile ?bench=perf: boosted-clock steady tok/s vs roofline

register-blocked GEMV lab: B=1/2/4/8/16 rows per group

profiler: fix read-only binding + dot8 batched-verify proxy + refined verdict

kernel limiter profiler: read-only vs ALU-intensity vs occupancy sweep

spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec

spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec

spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec

spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec

ternary GEMV kernel lab: V0/V1/V2 variants

roofline: batched passes + onSubmittedWorkDone for real bw

roofline: fill buffer non-zero + exceed cache for real VRAM bw

Upload roofline.html with huggingface_hub

Upload roofline.html with huggingface_hub

Upload roofline.html with huggingface_hub

Upload index.html with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload gpu-blake3.mjs with huggingface_hub

Upload qvac-gpu.js with huggingface_hub

Upload holo-gpu-device.mjs with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload index.html with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload gpu-blake3.mjs with huggingface_hub

Upload blake3-gpu-parallel.html with huggingface_hub

Upload index.html with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload holo-blake3.mjs with huggingface_hub

Upload blake3-gpu-test.html with huggingface_hub

Upload index.html with huggingface_hub

Upload qvac-gpu.js with huggingface_hub

Upload core/loader.js with huggingface_hub

Upload index.html with huggingface_hub

Upload qvac-gpu.js with huggingface_hub

Upload holo-load2bit.mjs with huggingface_hub

Upload folder using huggingface_hub

initial commit