search

ShuangXieIrene / ssds.pytorch

Repository for Single Shot MultiBox Detector and its variants, implemented with pytorch, python3.
MIT License
571 stars 166 forks source link

I train a new model,but test speed is too slow? #16

Open funnyuday opened 6 years ago

funnyuday commented 6 years ago

I down load the mobilenet_v1_ssd_voc_72.7, it's test speed faster(0.03s) than my new model(0.3s).

I can't find the reason,the configuration file is same as mobilenet_v1_ssd_voc_72.

I also train new mobilenetv2-ssd in same file, the speed is still very slow. Please give me some advice. Thank you very much!

foreverYoungGitHub commented 6 years ago

I just update the time_benchmark part and it will record the timing into a csv file. Would you mind to run the code again and share the csv file?

Best,

funnyuday commented 6 years ago

Thank you for your reply,I have solved this problem. My model spend more time on running NMS. I fine-tune the mobilenet_v1_ssd_voc_72.7.The speed is very fast.

nekulkarni commented 5 years ago

@funnyuday How did you finetune? I want to finetune the same model with my custom dataset as well, but I can't find out how to do that. Thanks in advance!

blueardour commented 5 years ago

@funnyuday I found a similar problem. The NMS is very very slow on my machine. Could you share what you have revised to improve the inference detection speed?

funnyuday commented 5 years ago

@blueardour I used the early training model to test.It need too many boxes to NMS.