search

blessedlex / Skynet-Zuckerberg-Edition

CYBR 4580-8950 IA Capstone Project
0 stars 0 forks source link

how can we make the most of stratified random sampling? #33

Closed blessedlex closed 3 years ago

blessedlex commented 3 years ago

Feed images to train the model from what’s left, the ones that aren’t used, run through the model after and see what’s collected. To get sense of false negative, false positive, true negative, etc.

What sample training sets do we need to get started? Let's try to figure what specific test models we want to use from the provided image sample libraries.

blessedlex commented 3 years ago

Using the COCO image library we will use a training model size of 5,000 total to create test models

blessedlex commented 3 years ago

Closing issue now that we have our training completed.