dsh-livebench-panel 0.2.1 → 0.2.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/lib/index.js +16 -0
- package/package.json +1 -1
package/lib/index.js
CHANGED
|
@@ -657,6 +657,13 @@ function apply(ctx) {
|
|
|
657
657
|
...process.env,
|
|
658
658
|
HF_HOME: join(layout.root, ".hf_cache"),
|
|
659
659
|
HF_HUB_DISABLE_SYMLINKS_WARNING: "1",
|
|
660
|
+
// 限制评测进程的数值库线程(numpy/pyarrow/MKL 默认吃满全核,
|
|
661
|
+
// 多模型并发时会把 node 事件循环饿死 → websocket 断连)。
|
|
662
|
+
OMP_NUM_THREADS: "1",
|
|
663
|
+
MKL_NUM_THREADS: "1",
|
|
664
|
+
OPENBLAS_NUM_THREADS: "1",
|
|
665
|
+
NUMEXPR_NUM_THREADS: "1",
|
|
666
|
+
RAYON_NUM_THREADS: "1",
|
|
660
667
|
// run_livebench.py shells out to bare `python`; without the venv on
|
|
661
668
|
// PATH that resolves to the system interpreter and dies on the first
|
|
662
669
|
// livebench import (shortuuid etc.).
|
|
@@ -694,6 +701,15 @@ function apply(ctx) {
|
|
|
694
701
|
env,
|
|
695
702
|
windowsHide: true,
|
|
696
703
|
});
|
|
704
|
+
// 降低评测进程优先级:评测不应与 DSH/浏览器争抢 CPU
|
|
705
|
+
try {
|
|
706
|
+
if (child.pid) {
|
|
707
|
+
spawn("powershell", [
|
|
708
|
+
"-NoProfile", "-Command",
|
|
709
|
+
`(Get-Process -Id ${child.pid}).PriorityClass = 'BelowNormal'`,
|
|
710
|
+
], { windowsHide: true, stdio: "ignore" }).on("error", () => {});
|
|
711
|
+
}
|
|
712
|
+
} catch { /* 优先级设置失败不影响评测 */ }
|
|
697
713
|
|
|
698
714
|
const record = {
|
|
699
715
|
id: `${Date.now()}-${Math.random().toString(36).slice(2, 6)}`,
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "dsh-livebench-panel",
|
|
3
|
-
"version": "0.2.
|
|
3
|
+
"version": "0.2.2",
|
|
4
4
|
"description": "DSH web plugin: a LiveBench tab in the Trajectory view (right of 对话/轨迹). Run LiveBench evaluations against every model configured in the DeepSeek Harness — pick provider/model, category, task, release and question range from dropdowns, watch progress, and read scores in place.",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"type": "module",
|