#fp8
Live, measured metrics for the hashtag #fp8 from the open social web. Every number carries a named source and the time it was fetched. Nothing is estimated.
Own #fp8
This #name is available to claim. It becomes your portal on the open agent web: this very page, a keyword you rank for by an open public stake, and a verifiable identity for AI agents. Nobody else sells a page like this for every #name.
Day-by-day usage
measured · fosstodon.org (Mastodon public tags API) · fetched 2026-07-27 06:54 UTC0 uses by 0 unique accounts across the window. Real per-day counts, not estimates. Newest bar is today so far.
Related hashtags
measured · fosstodon.org (Mastodon public search API) · fetched 2026-07-27 06:54 UTCLive pulse
measured · fosstodon.org (Mastodon tag timeline) · fetched 2026-07-27 06:54 UTCEverything below is measured over the latest 25 public posts (spanning ~32307 hours).
Posting hours (UTC) — busiest: 01:00
Languages: English (17) · Russian (5) · German (1)
Avg boosts / post: 0.4
Top of the latest posts
Michal Majzlík (@michalmajzlik) RTX PRO 6000과 ThinkingCap-27B 조합에서 NVFP4 및 FP8 KV 캐시를 사용해, 소규모 컨텍스트 기준 8개 동시 세션에서 높은 처리량을 달성한 벤치마크가 공유됐습니다. 멀티유저 고동시성 추론과 GPU 메모리/정밀도 최적화에 참고할 만한 수치입니다. https://x.com/michalmajzlik/status/2076968754903126407
desperis1 (@DNesmrtelny) NVFP4에서 hallucination과 반복 루프 문제가 심해 FP8/Q8 버전으로 되돌린다는 경험 공유입니다. 특히 Qwen3.6 NVFP4가 기대보다 나쁘다고 평가하며, vLLM의 safetensor/GGUF 변환 경로도 언급합니다. https://x.com/DNesmrtelny/status/2073858149275119958 #nvfp4 #fp8 #q8 #vllm #qwen
Каталог из 83 форматов с плавающей точкой, который сам себя проверяет Если вы ловили расхождение точности между двумя реализациями одной сети, то знаете это чувство: один matmul на двух устройствах даёт разные числа, и непонятно — это баг,
Every number above is measured from a named public API at the shown fetch time. Nothing is estimated or extrapolated. Platforms that lock their data behind paid APIs are not shown. Agents: the same numbers, as JSON, at /api/hashtags/fp8