dsh-livebench-panel 0.2.3 → 0.2.4
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 +4 -2
- package/package.json +1 -1
package/lib/index.js
CHANGED
|
@@ -25,7 +25,7 @@
|
|
|
25
25
|
* @module dsh-livebench-panel
|
|
26
26
|
*/
|
|
27
27
|
import { spawn } from "node:child_process";
|
|
28
|
-
import { readdirSync, readFileSync, existsSync, statSync, writeFileSync } from "node:fs";
|
|
28
|
+
import { readdirSync, readFileSync, existsSync, statSync, writeFileSync, mkdirSync } from "node:fs";
|
|
29
29
|
import { createRequire } from "node:module";
|
|
30
30
|
import { delimiter, dirname, join } from "node:path";
|
|
31
31
|
import { homedir } from "node:os";
|
|
@@ -365,7 +365,9 @@ function displayModelName(providerId, modelId) {
|
|
|
365
365
|
*/
|
|
366
366
|
function writeGeneratedModelConfig(layout, YAML, { displayName, modelId, reasoningEffort, protocol }) {
|
|
367
367
|
const configDir = join(layout.livebenchDir, "model", "model_configs");
|
|
368
|
-
|
|
368
|
+
// 每个模型一个独立配置文件：并发启动多模型时共享单文件会互相覆盖，
|
|
369
|
+
// 导致先启动的进程读到别人的配置 → anthropic/responses 路由失效 → 全 $ERROR$。
|
|
370
|
+
const target = join(configDir, "dsh_panel_generated__" + displayName.replace(/[^A-Za-z0-9._@-]/g, "_") + ".yaml");
|
|
369
371
|
void YAML;
|
|
370
372
|
const lines = [
|
|
371
373
|
"# Generated by dsh-livebench-panel — regenerated on every evaluation start.",
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "dsh-livebench-panel",
|
|
3
|
-
"version": "0.2.
|
|
3
|
+
"version": "0.2.4",
|
|
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",
|