Watson Visual Recognition Services的文档指出,该服务的成本为
$0.25/Training image
$0.004 for classification per image per custom class
$10 for storage per custom class per month因此,如果我有一个自定义分类器,每个分类器有1000个训练类,每个分类器有50个图像。那么成本就会是
$0.25 * 50000 = $12500 for initial training
$10 * 1000 = $10000 per month for storage
$4 per classification call if tested against all 1000 classes in the classifier我的理解正确吗?每通电话4美元似乎太高了。成本是每类(在本例中为1000)还是每个自定义分类器(在本例中为1)?
如果我后来添加了更多的训练图像(比如额外的500张图像),每个训练图像的$0.25是只对这些额外的图像收费($0.25 * 50500 = $125),还是$0.25 *50500= $12625?
https://stackoverflow.com/questions/38480586
复制相似问题