glm-5.3-flash-GGUF-1bit-dgx-spark cahlen · PARTIAL
Measured-optimal GLM-5.3-Flash GGUF (UD-IQ1_S) serving configuration for the NVIDIA DGX Spark / GB10 — agentic coding, MTP self-speculation, 128K context
github.com/cahlen/glm-5.3-flash-GGUF-1bit-dgx-spark · ★ 0 · Forks 0 · Size 452 KB
SUMMARY
Technologies 11
Scored 10
Observed 10
Practices 6
Evidence 12
Skips 1
COVERAGE
Analyzed 109 files · 28 commits · 0 API calls
TECHNOLOGIES & DEPTH
Shell LANGUAGE Depth 70
16 files · PRODUCTION
JSON LANGUAGE Depth 70
63 files · PRODUCTION
Markdown LANGUAGE Depth 70
4 files · PRODUCTION
TOML LANGUAGE Depth 70
1 files · PRODUCTION
YAML LANGUAGE Depth 70
1 files · PRODUCTION
Python LANGUAGE Depth 70
12 files · PRODUCTION
uv BUILD_TOOL Depth 80
1 files · CONFIGURATION
Poetry BUILD_TOOL Depth 80
1 files · CONFIGURATION
HTTPX LIBRARY Depth —
0 files · config only
Redis CACHE Depth 81
8 files · CONFIGURATION
pytest TESTING Depth 49
4 files · TEST
PRACTICES
documentation · observedautomated_tests · observedcontinuous_integration · observedcontainerization · absentlinting · observedformatting · absent
ACTIVITY & OWNERSHIP
First commit 2026-09-01
Last commit 2026-09-08
Active months 2
Commits 28