dotnet / dotnet/machinelearning-modelbuilder

Suggest feature: Object as Input Data instead of Link to location on Model Builder. Object detection. Predictor working slowly.

Open
#2,735 2 comments 0 reactions 0 assignees View on GitHub
Stale
Dominant language
Dockerfile
Stars
285
Forks
66
PR merge metrics
No merged PRs in 30d

Description

**Technical information:**
VS 2022
Model Builder: 17.17.0.2332602

I would like to be able to send an image object to the predictor instead of a link to the location. This would speed up image processing in real time, since it is not very competent and correct to save a frame to disk each time and send its location to the predictor. It is better to send him the object of the frame itself, for example, Bitmap.

The second suggestion: Reducing the time of data processing by the predictor. Each processing takes 5 seconds. Loading an 800x600 (100 kb) image takes more than 3-4 seconds (5 seconds according to my observation). I tried different sizes of images with different volumes, the result is the same. I tried to work with an SSD drive instead of a hard drive to get more speed, the result has not changed.
**My characteristics:**
CPU: i3 7020g 2.3 Ghz
RAM: 16 GB
Graphics card: Intel HD620 + nVidia MX130 2GB

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.