dsh-livebench-panel 0.2.4 → 0.2.6
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 +6 -1
- package/package.json +1 -1
package/lib/index.js
CHANGED
|
@@ -726,6 +726,9 @@ function apply(ctx) {
|
|
|
726
726
|
"--livebench-release-option", release,
|
|
727
727
|
"--max-tokens", String(asInt(body.maxTokens, 256, 32768, 32000)),
|
|
728
728
|
"--parallel-requests", String(asInt(body.parallel, 1, 8, 2)),
|
|
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) {
|
|
@@ -769,7 +772,9 @@ function apply(ctx) {
|
|
|
769
772
|
} catch { /* credential seam unavailable — fall through to env */ }
|
|
770
773
|
if (!key && process.env[provider.keyEnv]) key = process.env[provider.keyEnv];
|
|
771
774
|
}
|
|
772
|
-
if (provider.api === "openai-completions") {
|
|
775
|
+
if (provider.api === "openai-completions" || provider.api === "openai-responses" || provider.api === "openai_responses") {
|
|
776
|
+
// openai-responses 与 openai-completions 一样走 --api-base + LIVEBENCH_API_KEY
|
|
777
|
+
// （Responses 客户端同样从 api_dict 读 base_url/api_key）。
|
|
773
778
|
args.push("--api-base", provider.baseURL);
|
|
774
779
|
if (key) env.LIVEBENCH_API_KEY = key;
|
|
775
780
|
} else if (provider.api === "anthropic-messages") {
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "dsh-livebench-panel",
|
|
3
|
-
"version": "0.2.
|
|
3
|
+
"version": "0.2.6",
|
|
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",
|