nvidia-gb10

Reproducible refusal-subspace editing and TP=2 deployment of DeepSeek V4 Flash 0731 on two DGX Sparks

Native MiniMax-H3 video/audio inference in C — CUDA on Linux, Metal on Apple Silicon — with h3c studio, a self-hosted web UI.

DeepSeek-V4-Flash + DSpark speculative decoding on a pair of NVIDIA DGX Sparks (vLLM TP=2 over RoCE) — tuned recipe, overlays that halve multi-turn TTFT, contamination-guarded benchmarks, ops runbook

Benchmark for GB10 - Nvidia DGX Spark

Qwen3.8-Flash-Next NVFP4 + dense FP8, MTP K3 and C6 on one NVIDIA DGX Spark

One-command NVIDIA DGX Spark (GB10) cluster GPU monitoring with Grafana + Prometheus + DCGM + node_exporter + vLLM. Track GPU temperature, utilization, power, SM clock, memory, disk, network and LLM inference throughput in a pre-built dashboard. 一条命令搭建 DGX Spark 集群监控

Measured Muse Glimmer 30B recipe for NVIDIA DGX Spark, with llama.cpp, DFlash parity checks, vision, tools, and GB10 benchmarks.

Reproducible high-speed Qwen3.6-35B-A3B inference on NVIDIA GB10

Reclaim the 2046 MiB display reservation as ordinary RAM on a headless NVIDIA DGX Spark

Local LLM serving and measurement tools for Codex CLI on NVIDIA DGX Spark (GB10): a Responses API gateway over vLLM, a vLLM runtime manager for Qwen3.8-Flash-Next, an experimental llama.cpp path for Bonsai 2, and a research CUDA decoder.

DGX Spark & GB10 software: a private, fully local AI coding agent, chat and desktop app. One-command install, no cloud, no account, no telemetry. For NVIDIA DGX Spark, ASUS Ascent GX10, Dell Pro Max GB10, HP ZGX Nano, Lenovo ThinkStation PGX, MSI EdgeXpert, Acer GN100, Gigabyte AI TOP ATOM.

Private, on-premise AI workstation that turns a tumour variant file into an explainable, structure-guided shortlist of neoantigen research candidates. Runs fully offline on an HP ZGX Nano (NVIDIA GB10). Validated against 2,555 lab-tested peptides.

Headless MuJoCo on NVIDIA DGX Spark (GB10): EGL PNGs, Franka Panda, 6-D pinch (GRASP_OK), batched MJX/Warp. GPU beats one CPU Panda at 2048 (MJX) / 1024 (Warp) envs. Verified August-September 2026.

Auditable DeepSeek V4 Flash inference evidence on two NVIDIA GB10 systems

Three small, model-specific inference engines for NVIDIA DGX Spark.

Reproducible Qwen3.8-27B Q4/Q6 and llama.cpp MTP benchmarks on NVIDIA GB10

Production-ready local AI agent stack for NVIDIA DGX Spark / ASUS GB10. Gemma 4 31B NVFP4 + bge-m3 embeddings + OpenClaw gateway, one docker compose up.

Idle vs llm vs vl boot profiles on an NVIDIA DGX Spark (GB10): systemd user targets, linger, GPU free or Qwen3-VL-32B alone.

Atlas Inference Engine with Ling 3.0 Flash NVFP4 + MTP support for NVIDIA GB10 / DGX Spark

Empirical NVIDIA GB10 and SM121 microarchitecture characterization