Hugging Face's logo

HelloSun
/
Experiment_1

GGUF
conversational
Model card Files
xet
 Community
HelloSun commited on
Commit
75af4d6
·
verified ·
1 Parent(s): 18f561f

Upload opt/summary.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. opt/summary.md +3 -1
opt/summary.md CHANGED
@@ -41,4 +41,6 @@ llama-perplexity --kl-divergence, 8 chunks x 512 tokens (4096 positions), base =
41
  - Shared host: absolute numbers move +-15% between time windows; every comparison above was interleaved.
42
  - Under heavy neighbour load the 16-thread (full-quota) runs can stall because of cgroup throttling.
43
  - Remaining headroom: 137 GB/s achieved vs ~185 GB/s raw read. Prefill (pp) not optimized.
44
- - Source changes are NOT uploaded (data only); patch is kept locally.
 
 
 
41
  - Shared host: absolute numbers move +-15% between time windows; every comparison above was interleaved.
42
  - Under heavy neighbour load the 16-thread (full-quota) runs can stall because of cgroup throttling.
43
  - Remaining headroom: 137 GB/s achieved vs ~185 GB/s raw read. Prefill (pp) not optimized.
44
+ - The patch is published as openpangu-vl-cpu-vnni.patch in HelloSun/openPangu-VL-7B-Q4KM-gguf (opt-in: CPU_VNNI=1).
45
+ - Reproduced from scratch with the published build script (verify2_ab.jsonl): stock tg 15.1-15.4 vs VNNI 21.6-22.5 t/s
46
+ (busier host window, ratio ~1.44x).