jiazhizhong/higress
1
0
Fork 0
mirror of https://github.com/alibaba/higress.git synced 2026-06-09 20:57:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix error due to incompatible top_p range (#1250)

Browse Source
This commit is contained in:
Yang Beining
2024-08-26 12:28:43 +02:00
committed by GitHub
parent f5b8341f7f
commit 5a87031c0e
1 changed files with 12 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
plugins/wasm-go/extensions/ai-proxy/provider/qwen.go
View File

@@ -101,6 +101,18 @@ func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b
log.Errorf("Replace model error: %v", err) log.Errorf("Replace model error: %v", err)
return types.ActionContinue, err return types.ActionContinue, err
} }
// TODO: Temporary fix to clamp top_p value to the range [qwenTopPMin, qwenTopPMax].
if topPValue := gjson.GetBytes(body, "top_p"); topPValue.Exists() {
rawTopP := topPValue.Float()
scaledTopP := math.Max(qwenTopPMin, math.Min(rawTopP, qwenTopPMax))
newBody, err = sjson.SetBytes(newBody, "top_p", scaledTopP)
if err != nil {
log.Errorf("Failed to replace top_p: %v", err)
return types.ActionContinue, err
}
}
err = proxywasm.ReplaceHttpRequestBody(newBody) err = proxywasm.ReplaceHttpRequestBody(newBody)
if err != nil { if err != nil {
log.Errorf("Replace request body error: %v", err) log.Errorf("Replace request body error: %v", err)