CaliTrain

From Calimacil
Jump to: navigation, search

Description

This is the Calimotion learning machine (train.exe) which transform a log file made by the command MPU and generate the file hit.mod. This is a MS-Windows 64-bit version. You will need to have data sample under a sub-folder to generate the file hit.mod. After the file been generated, you can upload it on the Calimotion to get a new hit detection quality for the AI.

Download

You can Download the CaliTrain Windows 64bits version >> here <<

Syntax

Usage: train [-solver N] [-pre N] [-post N] [-mvavg N] [-threshold N] [-runs N] [-bias 0|1] [-o file] DATADIR

Parameters

[-solver N] Algorithm solver to apply for the machine learning, default value is (1) L2R_L2LOSS_SVC_DUAL
Solvers available (default 1)
  • 0 : L2R_LR
  • 1 : L2R_L2LOSS_SVC_DUAL
  • 2 : L2R_L2LOSS_SVC
  • 3 : L2R_L1LOSS_SVC_DUAL
  • 4 : MCSVM_CS (N/A)
  • 5 : L1R_L2LOSS_SVC
  • 6 : L1R_LR
  • 7 : L2R_LR_DUAL
  • 8 : N/A
  • 9 : N/A
  • 10 : N/A
  • 11 : L2R_L2LOSS_SVR
  • 12 : L2R_L2LOSS_SVR_DUAL
  • 13 : L2R_L1LOSS_SVR_DUAL
-pre [N] Number of data [N] samples to use before hit (default 20)
-post [N] Number of data [N] samples to use after hit (default 5)
-mvavg [N] Number of data samples to use for moving average filter (default 5)
-threshold [N] Peak detection threshold value (default 0.3)
-runs [N] Number [N] of times to train on data (default 10)
-bias [N] Use [N] a bias for training (default 0)

Examples

Simple commande Train.exe in MS-Windows

Create a hit.mod file in a MS-Windows Command line from data in folder /data

C:\Projects\CaliTrain>train ./data ok : 1 pre : 20 post : 5 bias : 0 mvavg : 5 threshold : 0.3 solver : 1 runs : 10 outfile : out.model Loading data from : ./data/christian_709.hit .. : 78 entries Loading data from : ./data/christian_709.nohit .. : 10 entries Loading data from : ./data/christian_709_2.nohit .. : 29 entries ............................**.**.*** optimization finished, #iter = 306 Objective value = -12.248396 nSV = 29 Correct = 35 out of 36 ...........................*..***.* optimization finished, #iter = 302 Objective value = -11.483025 nSV = 29 Correct = 34 out of 36 ..........................*....*.*.**.** optimization finished, #iter = 337 Objective value = -10.250934 nSV = 25 Correct = 31 out of 36 .........................*.** optimization finished, #iter = 268 Objective value = -9.069151 nSV = 26 Correct = 33 out of 36 .........................**.* optimization finished, #iter = 264 Objective value = -10.130809 nSV = 28 Correct = 32 out of 36 .......................** optimization finished, #iter = 238 Objective value = -8.396410 nSV = 21 Correct = 33 out of 36 ...........................***. optimization finished, #iter = 280 Objective value = -11.008693 nSV = 28 Correct = 33 out of 36 ..........................*...* optimization finished, #iter = 294 Objective value = -10.729508 nSV = 27 Correct = 32 out of 36 ............................*.**. optimization finished, #iter = 300 Objective value = -11.953233 nSV = 28 Correct = 36 out of 36 .........................*.* optimization finished, #iter = 262 Objective value = -11.870597 nSV = 27 Correct = 34 out of 36 ===================================================================* Results for each run...

0 : 0.972222 1 : 0.944444 2 : 0.861111 3 : 0.916667 4 : 0.888889 5 : 0.916667 6 : 0.916667 7 : 0.888889 8 : 1 9 : 0.944444


*

Average success rate : 0.925 ===================================================================*

...........................*.*.** optimization finished, #iter = 297 Objective value = -15.038237 nSV = 33 Correct = 0 out of 0

C:\Projects\CaliTrain>