Closed ambr89 closed 4 years ago
Hi, I want to ask the same/similar question. I trained my own dataset based on yolov3-tiny.cfg but with some minor tweaks: I only changed the number of classes
(1 in my case) and filters
( 18 in my case) accordingly. Do I need to change any code?
I ran commands per instructions and it seemed to finish successfully and a file yolo_v3.tflite
was generated. But I still want to ask this question to ensure I did it correctly.
PS: I've tried almost every project on github about converting yolov3 models to tensorflow (lite), and this is the only one that works for me (although pending on my further testing). Thanks!
@kuailink You only need to change the coco.names
file.
@peace195 Perfect! Yes, I did replace it with my content. Thanks for your repo again!
Hi!
What's happened if I've trained my tiny yolo v3 with a custom setting in .cfg file?
I've change width and height, I've change anchor and threshold.