R错误:无法分配大小为1.3GB的向量 [英] R error: cannot allocate vector of size 1.3GB
问题描述
大家好,
我们似乎遇到了同样的问题: https://social.msdn.microsoft.com/Forums/azure/en -US / 9183a2af-b8b1-4ef6-97a9-1bb2b5635092 / r-script-rtryeval-return-error-error-can-allocate-vector-of-size-714-mb?forum = MachineLearning
当尝试在模型上运行预测时,本地运行正常的R脚本会停在此处。不知道为什么R运行的实例无法访问这么多连续的内存,因为我知道它应该有50GB内存? (https://social.msdn.microsoft.com/Forums/azure/en-US/38625025-2e27-4a3c-a0c6-cda0755f1ea4/cannot-allocate-vector-of-size-37-gb?forum=MachineLearning)
我们可以在批处理执行模式下使用端点,但我们确实需要它在请求 - 响应模式下运行,我们不明白为什么这应该有所不同?
有人建议在请求 - 响应模式下使用端点时避免出现这些内存错误的解决方案吗?
干杯!
您好,
对不起您的不良体验感到抱歉。
我们希望就此问题与您进一步合作。您能否发送电子邮件至azcommunity@microsoft.com与我们联系?请向我们提供您的订阅ID和此主题的网址。
问候,
宇通
Hi everyone,
We seem to be having the same problem as this: https://social.msdn.microsoft.com/Forums/azure/en-US/9183a2af-b8b1-4ef6-97a9-1bb2b5635092/r-script-rtryeval-return-error-error-cannot-allocate-vector-of-size-714-mb?forum=MachineLearning
An R script that runs fine locally stalls here when trying to run predict on a model. Not sure why the instance of R running can't access this much contiguous memory as I understand it should have 50GB RAM? (https://social.msdn.microsoft.com/Forums/azure/en-US/38625025-2e27-4a3c-a0c6-cda0755f1ea4/cannot-allocate-vector-of-size-37-gb?forum=MachineLearning)
We can use the endpoint in batch execution mode, but we really need it to run in request-response mode, and we don't understand why this should make any difference?
Can anyone suggest a solution to avoid these memory errors when using the endpoint in request-response mode?
Cheers!
Hi,
Sorry for your bad experience.
We'd like to work with you further on this issue. Can you please email us at azcommunity@microsoft.com? Please provide us with your subscription ID and the URL for this thread.
Regards,
Yutong
这篇关于R错误:无法分配大小为1.3GB的向量的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!