###环境变量 GOOGLE_CLOUD_PROJECT=cryptcyber GOOGLE_CLOUD_LOCATION=global API_ENDPOINT=aiplatform.googleapis.com MODEL_ID="gemini-2.5-flash" GENERATE_CONTENT_API="generateContent" GOOGLE_API_KEY=AQ.Ab8RN6JIiaqs9I7GMOQBLvtzk0H-QlpcUSdyZJMwoIh3nRvPtw ####请求 curl -v -X POST -H "Content-Type: application/json" "https://${API_ENDPOINT}/v1/projects/${GOOGLE_CLOUD_PROJECT}/locations/${GOOGLE_CLOUD_LOCATION}/publishers/google/models/${MODEL_ID}:${GENERATE_CONTENT_API}?key=${GOOGLE_API_KEY}" -d @request.json [root@i-1 image]# cat request.json { "contents": [ { "role": "user", "parts": [ { "text": "How does AI work" } ] } ] , "generationConfig": { "temperature": 1 ,"maxOutputTokens": 32768 ,"responseModalities": ["TEXT"] ,"topP": 0.95 }, "safetySettings": [ { "category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF" }, { "category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF" }, { "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF" }, { "category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF" } ] }
24天前
错误信息: 所有令牌分组 default 下对于模型 gemini-3-pro-preview-11-2025-thinking 均无可用渠道，请更换分组尝试 (request id: 20251203171131956418773so5t1pCR) 堆栈信息: AI_APICallError: 所有令牌分组 default 下对于模型 gemini-3-pro-preview-11-2025-thinking 均无可用渠道，请更换分组尝试 (request id: 20251203171131956418773so5t1pCR) at file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/dist-mVY8hbYS.js:5132:11 at async postToApi (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/dist-mVY8hbYS.js:5028:24) at async OpenAICompatibleChatLanguageModel.doStream (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:241677:48) at async wrapStream (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:38808:31) at async fn (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:36993:17) at async file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:34371:19 at async _retryWithExponentialBackoff (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:34473:10) at async streamStep (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:36961:108) at async fn (file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:37240:5) at async file:///D:/cs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-Z-tJeJfQ.js:34371:19 错误原因: "[undefined]" 状态码: 503 请求路径: https://www.dmxapi.cn/v1/chat/completions 请求体: { "model": "gemini-3-pro-preview-11-2025-thinking", "user": "[undefined]", "max_tokens": "[undefined]", "temperature": 0.3, "top_p": "[undefined]", "frequency_penalty": "[undefined]", "presence_penalty": "[undefined]", "response_format": "[undefined]", "stop": "[undefined]", "seed": "[undefined]", "reasoning_effort": "[undefined]", "serviceTier": "[undefined]", "verbosity": "[undefined]", "messages": [ { "role": "user", "content": "Video Essay\n\nPlease follow the steps below to ensure that your video can be successfully uploaded.\n\n1. Please position yourself clearly in frame before recording. This video essay can be recorded using a laptop camera (recommended) or mobile phone.\n\n2. When you click \"Ready\" button, a question will appear beneath the video, and you will have 60 seconds to prepare a response before recording automatically begins. To begin recording earlier, click the \"Record\" button.\n\n3. Once recording has begun, you will have 120 seconds before the recording automatically stops. You may choose to stop the recording early (by clicking the \"Stop\" button) or may allow it to stop automatically at the end of the allotted time.\n\n4. IMPORTANT: After the \"Stop\" button is pressed or if the recording automatically stops, please replay the video in its entirety and review the video essay. The video will only be successfully uploaded if the buffer bar in the video has completely buffered. It has completely buffered when the scroll bar in the playback video does not stay at the extreme right and the un-played section of the scroll bar is a light grey colour. \n\n5. Please wait around 1 minute after replaying the video in its entirety before clicking on the \"continue\" button.\n\n*You may check if your video has been successfully uploaded when you navigate back to \"Video Essay\" tab and you are able to replay your video. \n\nTo begin the recording process, click the \"Ready\" button.\n\nNTU EEE的PhD线上申请的这个是什么意思？" }, { "role": "assistant", "content": "这段文字是新加坡南洋理工大学（NTU）电气与电子工程学院（EEE）博士（PhD）申请过程中，关于**“视频文书（Video Essay）”录制环节的具体操作指南**。\n\n简单来说，这是一个**类似于“视频面试”或“口语测试”**的环节，系统会随机出一个题目，你需要对着摄像头录制一段回答。\n\n以下是这段英文指南的详细中文解读和操作要点：\n\n### 核心流程解析\n\n1. **准备阶段 (Step 1):**\n * **设备要求：** 推荐使用带有摄像头的笔记本电脑，也可以用手机。\n * **站位/坐姿：** 确保你的人像清晰地出现在画面框中（建议背景整洁，光线充足，着装得体）。\n\n2. **审题与思考 (Step 2):**\n * 当你点击 **\"Ready\"（准备好了）** 按钮后，屏幕下方会出现一个**面试问题**。\n * 你有 **60秒** 的时间来思考如何回答。\n * 60秒倒计时结束后，录制会自动开始。如果你提前想好了，也可以手动点击 **\"Record\"（录制）** 按钮提前开始。\n\n3. **录制阶段 (Step 3):**\n * 录制开始后，你最多有 **120秒（2分钟）** 的时间来回答问题。\n * 你可以讲满2分钟让系统自动停止，也可以在回答完毕后手动点击 **\"Stop\"（停止）** 按钮结束录制。\n\n4. **回放与上传 (Step 4 - 非常重要):**\n * 录制结束后，**必须完整回放一遍你的视频**。\n * **缓冲条检查：** 这一步是为了确保视频上传成功。你需要观察播放进度条（Buffer bar）。只有当进度条完全缓冲（通常表现为未播放部分变成浅灰色，且滚动条不仅仅停留在最右边）才算成功。\n * 简单说就是：**一定要点开看一遍，确保视频能顺畅播完，没有卡顿或黑屏。**\n\n5. **最终确认 (Step 5):**\n * 在完整回放视频后，**请等待大约 1 分钟**，然后再点击 **\"Continue\"（继续）** 按钮。这是为了给服务器足够的时间接收文件。\n\n### 补充提示\n* **如何检查是否成功？** 你可以退回到 \"Video Essay\" 标签页，如果你能再次回放刚才录制的视频，就说明上传成功了。\n\n### 给你的建议（备考策略）\n\n既然是 PhD 申请的 Video Essay，题目通常不会太偏，建议提前准备以下几类话题的口语素材（控制在1.5-2分钟）：\n\n1. **自我介绍：** 你的背景、研究兴趣。\n2. **为什么选择 NTU/EEE？** 为什么选择这个项目？\n3. **研究经历：** 简述你做过的一个重要项目，遇到的困难及如何解决。\n4. **未来规划：** 博士毕业后的职业打算。\n5. **软技能类：** 你的优缺点、团队合作经历等。\n\n祝你申请顺利！", "tool_calls": "[undefined]" }, { "role": "user", "content": "我问了别人这个是什么，这是回答：\n今年第一年 之前都没有过\n如果是行为面试: 一般是问为什么申请这个学校这个项目\n如果是技术面试: CV 和PS中的项目内容要熟悉下,过往上的专业课程里的内容要熟悉 下 课,可能会被问到\n\n帮我全面的准备一下问题" } ], "tools": "[undefined]", "tool_choice": "[undefined]", "stream": true, "stream_options": { "include_usage": true } } 响应首部: { "content-length": "226", "content-type": "application/json; charset=utf-8", "date": "Wed, 03 Dec 2025 09:11:31 GMT", "document-policy": "include-js-call-stacks-in-crash-reports", "server": "nginx", "x-rixapi-request-id": "20251203171131956418773so5t1pCR" } 响应内容: { "error": { "code": "", "message": "所有令牌分组 default 下对于模型 gemini-3-pro-preview-11-2025-thinking 均无可用渠道，请更换分组尝试 (request id: 20251203171131956418773so5t1pCR)", "type": "rix_api_error" } } 数据: { "error": { "message": "所有令牌分组 default 下对于模型 gemini-3-pro-preview-11-2025-thinking 均无可用渠道，请更换分组尝试 (request id: 20251203171131956418773so5t1pCR)", "type": "rix_api_error", "code": "" } }
24天前
(<urllib3.connection.HTTPSConnection object at 0x7674a2546e30>, 'Connection to 192.168.10.100 timed out. (connect timeout=600)'))
24天前
":{"message":"Invalid URL (POST /v1/chat/completions:streamGenerateContent)","type":"invalid_request_error","code":""}}
24天前
Agent 执行失败: Failed to execute 'fetch' on 'Window': Failed to parse URL from https://www.dmxapi.cn:streamGenerateContent
24天前
Agent 执行失败: OpenAI 兼容 API 返回空响应内容
24天前
code: 400 - {'error': {'message': 'This response_format type is unavailable now', 'type': 'rix_api_error', 'param': '', 'code': 'invalid_request_error'}} 使用的 langchain
24天前
code: 400 - {'error': {'message': 'This response_format type is unavailable now', 'type': 'rix_api_error', 'param': '', 'code': 'invalid_request_error'}}
24天前
': {'message': 'This response_format type is unavailable now', 'type': 'rix_api_error', 'param': '', 'code': 'invalid_request_error'}}
24天前
': {'message': 'User error', 'type': 'video_generation_user_error', 'param': None, 'code': None}}
24天前
", "code": "stream_processing_failed" }
24天前
': {'message': 'User error', 'type': 'video_generation_user_error', 'param': None, 'code': None}}
24天前
This operation was aborted sending request'
24天前
code: 401 - {'error': {'message': 'Incorrect API key provided: sk-tc45z***************************************ThYA. You can find your API key at https://platform.openai.com/account/api-keys.', 'type': 'invalid_request_error', 'code': 'invalid_api_key', 'param': None}}
25天前
(Use `node --trace-warnings ...` to show where the warning was created) { role: 'assistant', content: '\nThe system is 3, and the file is 13K.' }
25天前