np.argmax() 与max() 区别

最新在nlp中想判断是不是训练过的类别，开始判断相似度，用了np.argmax()，但是debug的时候很奇怪，明明是才训练过的内容，为什么他是得出的是0，查了下，自己太傻了。

 np.argmax() 函数返回数组中最大值的索引。

max() 则是返回最大值

values = [9.99927630e-01, 8.32265116e-12, 1.53499887e-11, 1.65415892e-10, 8.35804443e-10, 1.70524309e-13, 6.60137006e-13, 1.93231280e-11, 3.45988649e-11, 3.45988649e-11, 1.74173837e-09, 4.35242268e-14, 2.24570815e-12, 7.78924844e-10, 9.08228885e-11, 1.00425861e-10, 3.36340489e-13, 4.19449144e-12, 6.26493980e-12, 2.13733349e-11, 1.94394145e-14, 4.52718829e-06, 9.40684709e-11, 1.50137876e-10, 1.38810104e-12, 6.99917958e-13, 9.20031205e-10, 6.68680520e-05, 1.13136486e-11, 8.21236535e-12, 8.29205301e-11, 4.77173195e-13, 3.19081582e-18, 4.82238799e-13, 7.81463600e-12, 9.36899723e-15, 1.32120079e-11, 1.66221328e-17, 3.61235754e-12, 4.31524051e-18, 3.28150429e-11, 7.70092299e-13, 1.67979932e-10, 2.51762309e-11, 1.74515095e-12, 1.61845336e-11, 1.16655358e-10, 2.54341434e-12, 3.94028130e-12, 2.20684628e-12, 1.04191542e-10, 1.64567234e-10, 1.75412582e-10, 7.28149232e-11, 1.24007543e-11, 1.58840339e-11, 2.97819865e-12, 9.68802333e-07, 3.46751442e-11, 2.53994046e-12, 1.71921992e-10, 1.11927732e-11, 4.11382078e-11, 1.48924042e-11, 4.72445884e-14, 5.97442906e-11, 2.53927357e-16]

max_value = max(values)
print(max_value) #0.99992763

values = [9.99927630e-01, 8.32265116e-12, 1.53499887e-11, 1.65415892e-10, 8.35804443e-10, 1.70524309e-13, 6.60137006e-13, 1.93231280e-11, 3.45988649e-11, 3.45988649e-11, 1.74173837e-09, 4.35242268e-14, 2.24570815e-12, 7.78924844e-10, 9.08228885e-11, 1.00425861e-10, 3.36340489e-13, 4.19449144e-12, 6.26493980e-12, 2.13733349e-11, 1.94394145e-14, 4.52718829e-06, 9.40684709e-11, 1.50137876e-10, 1.38810104e-12, 6.99917958e-13, 9.20031205e-10, 6.68680520e-05, 1.13136486e-11, 8.21236535e-12, 8.29205301e-11, 4.77173195e-13, 3.19081582e-18, 4.82238799e-13, 7.81463600e-12, 9.36899723e-15, 1.32120079e-11, 1.66221328e-17, 3.61235754e-12, 4.31524051e-18, 3.28150429e-11, 7.70092299e-13, 1.67979932e-10, 2.51762309e-11, 1.74515095e-12, 1.61845336e-11, 1.16655358e-10, 2.54341434e-12, 3.94028130e-12, 2.20684628e-12, 1.04191542e-10, 1.64567234e-10, 1.75412582e-10, 7.28149232e-11, 1.24007543e-11, 1.58840339e-11, 2.97819865e-12, 9.68802333e-07, 3.46751442e-11, 2.53994046e-12, 1.71921992e-10, 1.11927732e-11, 4.11382078e-11, 1.48924042e-11, 4.72445884e-14, 5.97442906e-11, 2.53927357e-16]

max_value = np.argmax(values)
print(max_value) #0

要不是查一下，我还一直在想不通为什么相似度最大是0.哈哈哈哈