I have a buffer of float16 elements in the memory. I read this buffer to a list but since I use np.ctypeslib I can't read it as float16. So I read it as uint16 and now I want to represent it as float16. Simple convert by list.astype(float16) won't help. I found this table https://gist.github.com/gregcotten/9911bda086c5850cb513b2980cca7648 and according to that table it works. I want to write a python function for convention unint16_as_float16 but don't know what are the rules. I appreciate your help. Continue reading...
