search

SkyWa7ch3r / ImageSegmentation

This project is a part of the Pawsey Summer Internship where I will do test multiple semantic segmentation algorithms and models on their training and inference time. There will also (given time) be experimentation with Panoptic Segmentation which combines semantic and instance segmentation together.
7 stars 4 forks source link

why my FPS is about ~14? #2

Open amiraliAI opened 3 years ago

amiraliAI commented 3 years ago

Hello Mr Joel I used your fast-scnn model for segmentation task and trained it. But in inference it gives 14 frame per second on gtx 1080 ti tensorflow2 and gpu. do you know what is wrong? input shape :512,512,3

SkyWa7ch3r commented 2 years ago

Hello Mr Joel I used your fast-scnn model for segmentation task and trained it. But in inference it gives 14 frame per second on gtx 1080 ti tensorflow2 and gpu. do you know what is wrong? input shape :512,512,3

Hi Sorry for a really delayed reply...I'm sure you've totally forgotten about it. I'm really not sure there, maybe it's the way the FastSCNN model upscale the content, also you may need to use a mixed precision to get the best inference FPS.

I'm not entirely sure what is wrong. Did you ever figure it out?