Public community directory addendum

Observed 2026-10-13. Publication dates remain unknown; these are navigation and methodology sources, not local benchmark results. Local Inference Lab Discord was reviewed read-only; private messages, channel links and account information are excluded.

Source	Evidence and relevance	How to use it
Local Inference Lab public organization	community-maintainer-directory; SM120 model/runtime research; repositories vary in topology	Follow public changes; do not count linked projects as independent corroboration.
RTX PRO 6000 Blackwell field wiki	maintainer-runbook-and-field-measurements; SM120 PCIe; most reported measurements use 4, 8 or 16 GPUs	Locate current model and kernel runbooks; transfer TP1/TP2 figures only with matched controls.
Local Inference Lab HF organization	artifact-publisher-directory; Model variants/QAD/CSF and immutable artifact discovery	Verify exact revision and architecture; organization metadata is not local quality evidence.
Blackwell runtime image builds	runtime-maintainer-source; SGLang/vLLM SM120 CUDA 13.2 images	Pin source/image and native dispatch before using a reported fix.
Local Inference Lab throughput harness	benchmark-harness-source; SGLang/vLLM decode context/concurrency sweeps	Throughput method lead; not independent repository-repair evidence.
Local Inference Lab quant toolkit	quant-maintainer-source-directory; Exact current support unverified in this addendum	Discover quant tooling; do not imply selected-model/runtime compatibility.
Local Inference Lab newcomer and Discord support guide	community-methodology; Exact image/model/topology/client/server evidence expected	Use when learning from Discord; preserve logs/settings and isolate one variable.
Local Inference Lab archived b12x	historical-kernel-source; Organization declares development moved to FlashInfer	Historical recipe lead; use current FlashInfer migration rather than treating archive as latest kernel.