search

sacmehta / ESPNet

ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation
https://sacmehta.github.io/ESPNet/
MIT License
541 stars 112 forks source link

Could you let me know the values in graph or indicate the exact numbers in your paper? #1

Closed EthanCalvin closed 6 years ago

EthanCalvin commented 6 years ago

Hi, I read your paper and want to know exact values in your graph in Figures 5,6,7. Could you let me know the values in graph or indicate the exact numbers in your paper?

sacmehta commented 6 years ago

See Table 2 for exact values of our module ESP.

Sent from my iPhone

On Mar 30, 2018, at 3:23 AM, EthanCalvin notifications@github.com wrote:

Hi, I read your paper and want to know exact values in your graph in Figures 5,6,7. Could you let me know the values in graph or indicate the exact numbers in your paper?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

sacmehta commented 6 years ago

Table 2 contains the accuracy and model size. Other values are hardware deprendent and you should be able to reproduce them with the source code.

EthanCalvin commented 6 years ago

Thank you for your quick reply, I need to reproduce other values with my hardware. but If you don't mind could I ask inference time for each model?

sacmehta commented 6 years ago

As I mentioned, these values are hardware dependent. If you have exact hardware (e.g. laptop) as listed in Appendix, you should get 112 fps for a resolution of 1024x512.

You need to reproduce values for other models yourself. Please note that we didn’t use CRF with any models while computing inference speed, so FPS will drop further if you are using CRF.

Please use this space for any code related issues.