咨询 claude 429

[user135 (ungzs)]

claude 报 429 了, 是不是号被扬了啊?

[
    {
        "error": {
            "code": 429,
            "message": "Quota exceeded for 
aiplatform.googleapis.com/online_prediction_requests_per_base_model 
with base model: anthropic-claude-3-5-sonnet. 
Please submit a quota increase request. 
https://cloud.google.com/vertex-ai/docs/generative-ai/quotas-genai.",
            "status": "RESOURCE_EXHAUSTED"
        }
    }
]

报的好像是这个限额, 没有放量了

地址 https://console.cloud.google.com/iam-admin/quotas?hl=zh-cn&project=artful-timing-429411-m0&pageState=(%22allQuotasTable%22:(%22f%22:%22%255B%257B_22k_22_3A_22%25E7%25BB%25B4%25E5%25BA%25A6%25EF%25BC%2588%25E4%25BE%258B%25E5%25A6%2582%25E4%25BD%258D%25E7%25BD%25AE%25EF%25BC%2589_22_2C_22t_22_3A10_2C_22v_22_3A_22_5C_22base_model_3Aanthropic-claude-3-5-sonnet_5C_22_22_2C_22s_22_3Atrue_2C_22i_22_3A_22displayDimensions_22%257D%255D%22))

但是其实还有额度的

[Qiner (林黛玉倒拔垂杨柳)]

跟客服对线一整天了，官方政策，已经撤了除 Opus 外的所有 Claude 配额。（Opus 还能秒提额，但后续是否会回收不知道）

[bbb]

现在配额全变0了

[bbb]

该来的还是来了

[handsome (大帅哥)]

gemini用用算了

[user135 (ungzs)]

看来地主家被薅疼了

[kill (悟空)]

客服最后咋说的，是不管什么账号claude sonnet3.5的配额都置为0了吗。
也不知道谷歌是出于什么考虑的，实在不行的话换aws平台了

[printfff]

aws现在都要水电单，3.5和op都要人工申请才行

[kill (悟空)]

aws买老号问题应该不大吧 我看网上有卖什么32v之类的老号

[printfff]

那玩意不是老号，而是你这个账号在创建时候能够开启的 CPU 数量

[printfff]

等价于你这个账号的风控高不高而已

[kill (悟空)]

标的是老号，能开claude模型的
我在隔壁nodeseek看到的

[printfff]

你确定是3.5和 op ？这两个都需要人工申请，其它模型的话不需要

[kill (悟空)]

你这样说的话我就不太确定了

[Qiner (林黛玉倒拔垂杨柳)]

反正就是说工程师回复不能提额，说要去用什么「预支吞吐量」或「动态共享配额」但这俩的可选参数都只有谷歌自家的 Gemini 那些完全没有 Claude 相关的。

[kill (悟空)]

Hello Alan,

I hope that you are doing well today.

This email is an update about your request for the project xxxxxx. Upon checking, your request requires the assistance of your Account Team/Google Sales Representative. Hence, we are unable to find your Account Team/Google Sales Representative. If you have one, please share their information by responding to this email. If you don't have one, please contact our Sales Team by going to this link: https://cloud.google.com/contact/.

Additionally, our Engineers mentioned that Claude models are now available through Dynamic Shared Quota and it eliminates the need for filing a quota increase request. Additionally for production workloads, we recommend utilizing the Provisioned Throughput by going to this link: https://cloud.google.com/vertex-ai/generative-ai/docs/provisioned-throughput.

Alternatively, you can deploy the Claude 3 model directly from the Model Garden to your own Vertex AI endpoint. To explore this option, visit the Model Garden within the Vertex AI console and search for "Claude" on this link: https://console.cloud.google.com/vertex-ai/model-garden.

Please let us know your thoughts about this. We will be leaving this support case open for 3 business days while waiting for your response.

Sincerely, 
Genesis on behalf of the Google Cloud Platform Team

我的是让我联系他们的销售，说claude有动态配额，但是没搞明白它这个动态配额是指什么

[Qiner (林黛玉倒拔垂杨柳)]

一样，动态共享配额、预配的吞吐量 都是说这里两个东西，后者新建根本就没有 Claude 相关，前者说什么都不用改自动支持，然而 429。

[kill (悟空)]

是的 然后我给他们回复了封邮件。把配额截图发他们了，目前还没收到回复。
现在美国时间是晚上十一点多了，估计要明天才能收到邮件了

[GPT-4]

如果以前有用过GD就知道谷歌的客服一点也不靠谱，每个客服的回复可能都不一样

[user135 (ungzs)]

我这边看 opus 也没有

准备切 gemini 用用得了