Spaces:
Running
Running
Commit History
-
Remove orphaned old-layout files (_shared vendor dup + old diagnostics)
Verified signature666a78d -
Unified feature-complete build: clause-streamed voice + 260-char bubbles + PWA/offline SW + self-contained vendor
Verified signature89dc30e
-
Q: canonical standalone chat rebuilt from apps/q engine (warm-KV persona-prefix reuse)
Verified signature63158fe
-
keep-warm pulse: hold GPU boost clock while typing so casual chat decodes at boosted rate
Verified signature70592b3 -
warmup uses decode() to precompile batched pipelines (cut first-msg TTFT)
Verified signaturee4b6a8f -
fast first turn: warmup boosts clock + primes system-prompt KV (12s TTFT to ~0.5s) + static grounded greeting
Verified signature7efaf4b -
discrete-GPU validation ?bench=discrete: bandwidth + live + spec-flip in one page
Verified signaturedabcb0d -
decode(): report live-path GPU ms/tok for clean CPU/GPU split
Verified signatured2afec3 -
decode(): report live-path GPU ms/tok for clean CPU/GPU split
Verified signature8f34f10 -
per-pass GPU trace ?bench=trace: name the non-weight overhead
Verified signature770541b -
per-pass GPU trace ?bench=trace: name the non-weight overhead
Verified signature31a05bf -
spec bench: warmup + short prompt + 192-tok decode-dominated (fix prefill confound)
Verified signature439999f -
live decode profile ?bench=perf: boosted-clock steady tok/s vs roofline
Verified signaturee74320a -
register-blocked GEMV lab: B=1/2/4/8/16 rows per group
Verified signature3653faa -
profiler: fix read-only binding + dot8 batched-verify proxy + refined verdict
Verified signaturebcdc5b6 -
kernel limiter profiler: read-only vs ALU-intensity vs occupancy sweep
Verified signaturea24d0c6 -
spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec
Verified signature7406205 -
spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec
Verified signatured4f3433 -
spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec
Verified signature459725d -
spec-decode for BitNet: subNorm+f32-KV batched verify, ?spec + ?bench=spec
Verified signature57800b6 -
ternary GEMV kernel lab: V0/V1/V2 variants
Verified signaturebbf8a0a -
roofline: batched passes + onSubmittedWorkDone for real bw
Verified signaturefa278c0 -
roofline: fill buffer non-zero + exceed cache for real VRAM bw
Verified signatureac2b557 -
Upload roofline.html with huggingface_hub
Verified signatured882c7e -
Upload roofline.html with huggingface_hub
Verified signature082923c -
Upload roofline.html with huggingface_hub
Verified signatured88b0fc -
Upload index.html with huggingface_hub
Verified signatureb50807f -
Upload holo-load2bit.mjs with huggingface_hub
Verified signaturef789de8 -
Upload holo-load2bit.mjs with huggingface_hub
Verified signatureb5ae11e -
Upload gpu-blake3.mjs with huggingface_hub
Verified signature4e8cc03 -
Upload qvac-gpu.js with huggingface_hub
Verified signature586e17f -
Upload holo-gpu-device.mjs with huggingface_hub
Verified signaturedf29783 -
Upload holo-load2bit.mjs with huggingface_hub
Verified signature139e306 -
Upload holo-load2bit.mjs with huggingface_hub
Verified signature81c5d37 -
Upload index.html with huggingface_hub
Verified signature1185393 -
Upload holo-load2bit.mjs with huggingface_hub
Verified signature5b034d2 -
Upload gpu-blake3.mjs with huggingface_hub
Verified signaturee020433 -
Upload blake3-gpu-parallel.html with huggingface_hub
Verified signaturea3c9d6f -
Upload index.html with huggingface_hub
Verified signatureabdc36e -
Upload holo-load2bit.mjs with huggingface_hub
Verified signature0a2d0d8 -
Upload holo-blake3.mjs with huggingface_hub
Verified signature7b5c4b3 -
Upload blake3-gpu-test.html with huggingface_hub
Verified signaturea349783 -
Upload index.html with huggingface_hub
Verified signaturecb2d355 -
Upload qvac-gpu.js with huggingface_hub
Verified signature78a757e -
Upload core/loader.js with huggingface_hub
Verified signature484edea -
Upload index.html with huggingface_hub
Verified signature7d0b228 -
Upload qvac-gpu.js with huggingface_hub
Verified signature813e8d8 -
Upload holo-load2bit.mjs with huggingface_hub
Verified signaturedfdc138 -
Upload folder using huggingface_hub
Verified signature3365e13
-
initial commit
Verified signature3abea9c