Daniel Konegen / MNIST_example

Dependencies:   mbed-os

Embed: (wiki syntax)

« Back to documentation index

Show/hide line numbers constants.h Source File

constants.h

00001 #ifndef TENSORFLOW_LITE_CONSTANTS_H_
00002 #define TENSORFLOW_LITE_CONSTANTS_H_
00003 
00004 // This constant represents the range of x values our model was trained on,
00005 // which is from 0 to (2 * Pi). We approximate Pi to avoid requiring additional
00006 // libraries.
00007 const float kXrange = 2.f * 3.14159265359f;
00008 
00009 // This constant determines the number of inferences to perform across the range
00010 // of x values defined above. Since each inference takes time, the higher this
00011 // number, the more time it will take to run through the entire range. The value
00012 // of this constant can be tuned so that one full cycle takes a desired amount
00013 // of time. Since different devices take different amounts of time to perform
00014 // inference, this value should be defined per-device.
00015 extern const int kInferencesPerCycle;
00016 
00017 #endif  // TENSORFLOW_LITE_CONSTANTS_H_