Deep Learning - How to use inputs on learning model

#1

I dont knwo how to use my input vector with values inside a deep learning model.

Imagine that I developed a library with many fuctions that have versions inside github. I have the analysis of wich function and the runtime that each spend to run:
• Previous Version of functions with its runtime
• Current Version of functions with its runtime
• Difference between the time (previou runtime - currente runtime)
• Name of function

I want to predict which function makes my program fail. For Example:

GIT PREVIOUS VERSION
Total runtime = 100s

Function A = 40
Function B = 50
Function C = 10

GIT CURRENT VERSION
Total runtime = 110s

Function A = 0s
Function B = 0s
Function C = 20s
New Function Z = 39s
New Function X = 52s

In the example above, I have the following analysis:
DIFFERENCE RUNTIME: PREVIOUS – CURRENT
Function A = 40 - 0 = 40s
Function B = 50 - 0 = 50s
Function C = 10 - 20 = -10s
Function Z = 0 - 39 = -39s
Function X = 0 - 52 = -52s

Using the DIFFERENCE RUNTIME I can estimate that the problem are the function C, because the function Z and function X are equivalence with function A and Function B.

wIth Python I have the vector that I use as input to predict wich function has the problem:
• X_diff
• X_prev
• X_curr

#2

Hey @leoengufmg, the question seems ambiguous and is not understandable in this current state. What daaset you have with your self. Morevover neither it seems regression problem nor classification as if for now, Plz be more clear.

#3

according to the runtime that each function takes in each commit, I would like to predict the functions that probably have a problem. On code that I shared, I create a list of weight that are used to make my criteria of which are wrong and which are not: https://ide.codingblocks.com/s/215204

#4

so with the weight and the vector runtimes inputs, I would like to predict the failures

#5

Hey @leoengufmg, what is your training data ?

#6

I tried to make the training model the same as Prateek Narang in the class, but my doubt is this, I didnt know how to use this inputs and weight inside a model or how to create the training model with my requirements.

#7

Hey @leoengufmg, first we need to identify the type of problem, we can’t just implement linear regression directly, so first of all an individual needs to be clear, what’s the problem. So for that, you need to study the training data, that’s what i am asking again for the third time, what is your training data file, comprising ?

#8

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.