dsh-livebench-panel 0.2.5 → 0.2.7
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 +7 -4
- package/package.json +1 -1
package/lib/index.js
CHANGED
|
@@ -725,7 +725,10 @@ function apply(ctx) {
|
|
|
725
725
|
...(benchNames !== null ? benchNames : [benchParts.join("/")]),
|
|
726
726
|
"--livebench-release-option", release,
|
|
727
727
|
"--max-tokens", String(asInt(body.maxTokens, 256, 32768, 32000)),
|
|
728
|
-
"--parallel-requests", String(asInt(body.parallel, 1, 8,
|
|
728
|
+
"--parallel-requests", String(asInt(body.parallel, 1, 8, 1)),
|
|
729
|
+
// 流式：中转网关（Cloudflare 等）对非流式请求有 ~100s 超时（524），
|
|
730
|
+
// 高推理强度模型思考数分钟必然超时。流式保持字节流动可规避。
|
|
731
|
+
"--stream",
|
|
729
732
|
"--mode", "single",
|
|
730
733
|
];
|
|
731
734
|
if (body.begin !== undefined && body.begin !== null && `${body.begin}`.length > 0) {
|
|
@@ -828,18 +831,18 @@ function apply(ctx) {
|
|
|
828
831
|
// 自动补跑：正常结束但存在 $ERROR$ 答案时，用 --resume --retry-failures
|
|
829
832
|
// 只重跑失败题（最多 2 轮）。统计改为异步流式（同步版曾阻塞事件循环，
|
|
830
833
|
// 且函数缺失会在 close 回调抛未捕获异常 → DSH 假死断连）。
|
|
831
|
-
if (record.exitCode === 0 && record.autoRetries <
|
|
834
|
+
if (record.exitCode === 0 && record.autoRetries < 3 && record.body) {
|
|
832
835
|
(async () => {
|
|
833
836
|
try {
|
|
834
837
|
const errors = await countErrorAnswersAsync(layout.dataDir, record.displayName, record.body);
|
|
835
838
|
if (errors > 0) {
|
|
836
|
-
record.log.push(`[自动补跑] 检测到 ${errors} 条失败答案，自动重试（第 ${record.autoRetries + 1}/
|
|
839
|
+
record.log.push(`[自动补跑] 检测到 ${errors} 条失败答案，自动重试（第 ${record.autoRetries + 1}/3 轮）`);
|
|
837
840
|
const retryBody = { ...record.body, resume: true, retryFailures: true };
|
|
838
841
|
setTimeout(() => {
|
|
839
842
|
startRun(retryBody, record.autoRetries + 1).catch((e) => {
|
|
840
843
|
appendLog(record, `[自动补跑启动失败] ${e.message}`);
|
|
841
844
|
});
|
|
842
|
-
},
|
|
845
|
+
}, 15000);
|
|
843
846
|
}
|
|
844
847
|
} catch (e) {
|
|
845
848
|
record.log.push(`[自动补跑检查失败] ${e.message}`);
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "dsh-livebench-panel",
|
|
3
|
-
"version": "0.2.
|
|
3
|
+
"version": "0.2.7",
|
|
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",
|