Pixel ML

Community-tested recipes, diagnostics, and benchmarks for CMP 170HX AI workloads

GLM-5.3-Flash on CMP 170HX (SM80): every attempt - NVFP4, AWQ, GPTQ, EXL3, FP8/BF16 - with static-fit math, runtime compatibility, and reproducible evidence

Community-tested recipes, diagnostics, and benchmarks for NVIDIA DGX Spark workloads

Reproducible SGLang NVFP4 deployment of Qwen3.8-Flash-Next across two NVIDIA DGX Spark systems.

DeepSeek-V4-Flash-Vision-Exp on 4x CMP 170HX (SM80): startup, benchmark evidence, and reproducibility

DeepSeek-V4-Flash-Vision-Exp on 2x DGX Spark (GB10) — fit/startup/performance evidence and verdicts

Live OpenAI-compatible parallel benchmark capture, dashboard, and HyperFrames publishing toolkit.

Qwen3.8-Flash-Next on CMP 170HX (SM80): static-fit math, runtime compatibility, and evaluation harness for AWQ INT4 and future variants

FastH3 4-step text-to-audio-video (FastVideo BF16 baseline + Kijai INT8 convrot experimental) on 4x CMP 170HX (SM80): compatibility, static memory fit, benchmark harness, and reproducible evidence