gufo-org / gufo

The Open Source Strix Halo AI engine. Qwen 3.8 Flash Next Q4_K_XL: 1,801.01 pp, 73.17 tg single user, 194.37 tok/s 8 users; Qwen 3.8 27B Q4_K_XL: 656.33pp, 70.56tg tok/s single user with DFlash2

Luce-Org / lucebox

LLM speculative inference server for heterogeneous hardware & consumer GPUs

jpezzulli / sglang-rtxpro6000

Optimized SGLang runtime for Qwen3.8-27B FP8 with DFlash2 and Qwen3.8 Flash-Next NVFP4 with FR-Spec on one NVIDIA RTX PRO 6000 Blackwell 96 GB GPU (SM120): 524K context, HiCache and NIXL.

Tencent / AngelSlim

Model compression toolkit engineered for enhanced usability, comprehensiveness, and efficiency.

ARahim3 / mlx-dspark

Up to 4× faster LLM decoding on Apple Silicon, lossless. Native MLX port of DeepSeek's DSpark & z-lab's DFlash speculative decoding — Gemma-4, Qwen3.8, Muse-Glimmer, Nemotron, LFM2.5, Ornith-1.0, ternary Bonsai-27B.

Tencent / AngelSpec

A unified, torch-native training framework for MTP and block-parallel speculative decoding.

tonyd2wild / GLM-5.3-Flash-NVFP4-DFlash2-2x-DGX-Spark

GLM-5.3-Flash (NVFP4) on 2x NVIDIA DGX Spark - vLLM TP2, 262K context, MTP. World-first deploy recipe: 7 day-0 bugs found and fixed, patched sm121 image, probes and full report.

MiaAI-Lab / Qwen3.8-27B-DFlash2-EXL3-5.0bpw

Deployment kit: Qwen3.8-27B EXL3 3.5bpw target + DFlash2 EXL3 5.0bpw speculative draft — launcher, config, model cards

alexellis / glm-5.3-flash-4x-dgx-spark-switchless

GLM-5.3-Flash (NVFP4) at TP4 across 4x DGX Spark via a switchless RoCE ring + DFlash2 — reproducible recipe

kindlingai / glm-5.3-flash-gx10

GLM-5.3-Flash (NVFP4) at TP=4 across 4x ASUS GX10 / GB10 - vLLM + mentat instead of Ray, 524k context, DFlash2. Minimal working recipe.

styles01 / sparkrun-recipes

Custom inference recipes for NVIDIA DGX Spark — Qwen 122B DFlash hybrid

headpiece747 / ninfer-5090-windows

Native Windows port of NInfer engine for RTX 5090. Features Qwen3.8-27B with QUASAR and NInfer models, MTP/DFlash2 with vision and 262,144 context

Reederey87 / glm53-flash-exl3-2x-dgx-spark

GLM-5.3-Flash EXL3 (320B MoE) on 2x NVIDIA DGX Spark — production serving kit, 1M context, 97%+ multi-session prefix caching, DFlash2 spec decode

0xBakeer / Qwen3.8-27B-FP8-on-a-single-DGX-Spark

Serving Qwen3.8-27B-FP8 on a single DGX Spark (GB10): 7.88 to 58.5 tok/s single-stream from decode strategy alone, weights untouched. Speculative decoding and prefix caching benchmarked, plus DFlash 2 — the only Qwen3.8-27B build that can serve it under vLLM.

gitcommit90 / glm-5.3-one-spark

GLM-5.3-Flash EXL3 2.05 on one DGX Spark — TP1 vLLM, DFlash2 K5, concurrency, 262K context, and published fidelity measurements

tonyd2wild / MiMo-V2.6-Flash-DGX-Spark-Recipe

MiMo-V2.6-Flash-RL on 2x NVIDIA DGX Spark (GB10): vLLM TP2 + DFlash + vision/audio, fp8 KV 300K, four fixes for the stock image, measured baselines

1115714829 / llama.cpp-v100

llama.cpp optimized for NVIDIA Tesla V100 (SM70): large MoE Qwen3.8-Flash-Next with MTP (prefill 3464 tok/s at 200K context, 6x V100-16GB) and Qwen3.8-27B with DFlash2 (decode up to 225 tok/s); 2-6 GPUs, 262K-512K context, multimodal, concurrent serving.

Entrpi / glm-5.3-flash-exl3-2x-spark

GLM-5.3-Flash (EXL3 4bpw + DFlash2) on 2x NVIDIA DGX Spark: one-shot installer, 33-74 tok/s c1, 1.3M+ context, vision

seanyourhighness / vllm-sm12x-nvfp4-dflash2

All-NVFP4 vLLM + DFlash2 K7 for Blackwell (SM120/SM121): NVFP4 target/draft/KV, 262K context, optional CPU vision sidecar.

WildFlash1st / cmp90hx-unlock

Unlocking NVIDIA CMP 90HX (GA102): kernel 6.12

knapcio / GLM-5.3-Flash-4x-DGX-Spark-TP4

GLM-5.3-Flash NVFP4 on 4x NVIDIA DGX Spark: vLLM TP4, DFlash2 adaptive draft, measured FP8 vs NVFP4 vs EXL3 vs SGLang

pangoleen / qwen3.8-27b-dgx-spark-dflash2

Qwen3.8-27B on one DGX Spark: SGLang + DFlash2 recipe, engine image, benchmarks, data and charts. By @redp314

Geramy / LSE

HRX, LOOM, Mac, and AMD running top speed kernels with a self optimizing engine.

DocPang / qwen38-k100ai-int8-optimization

Qwen3.8-27B INT8/W8A8 optimization for Hygon K100AI: SGLang, TP1/TP4, DFlash2 and long-context Agent workloads

WildFlash1st / cmp90hx-unlock-for-windows

Open-source EFI bootloader that unlocks compute + render of NVIDIA CMP 90HX (GA102) before Windows boots. Compute/AI focused: this mining card is a poor gaming GPU — Gen1 x16, no display outputs, patched driver required.

Weschera / Qwen3.8-27B-NVFP4-DFlash2-DGX-Spark

Reproducible Qwen3.8-27B NVFP4 + DFlash2 recipe and qualification evidence for one NVIDIA DGX Spark

jnardiello / GLM-5.3-Flash-FP8-4-DGX-Spark-Switchless

GLM-5.3-Flash (FP8) + DFlash2 speculative decoding on 4x NVIDIA DGX Spark / ASUS GX10: vLLM TP4, 256K context, switchless ConnectX-7 ring with patched NCCL

darkdatter / gb10-repo

Qwen3.8-27B on DGX Spark (GB10): SGLang + NVFP4 + DFlash2. Measured recipe, benchmark harness, and the traps that cost real time.

r0b0tlab / qwen38-exl3-dflash2

Qwen3.8-27B EXL3 (4.00 bpw) + DFlash2 speculative decoding for ExLlamaV3, validated at 262k context on a 24 GB RTX 3090

satellitedown / fafstmobel-cinference

One-menu RTX 5090 installer for fafstmobel with Cinference, DFlash2, K8V4, and vision.