natanielruiz / dockerface Goto Github PK

I get this error when running the make -j20 line while trying to compile Caffe. I am using an AWS GPU server with the new NVIDIA Voltas. Any hint how to fix this error?

Is CPU version availabe ? And can use it for live camera feed ?

Thx

For getting more accuracy in faster-rcnn , which parameters i have to tune (tuning parameters). please anybody provide the solution for that problem, actually my model is working but the problem is getting less accuracy and some miss classifications are happens , so please kindly provide solution for getting high accuracy in the model.

Thanking you!

Hi Nataniel,

In your instructions in the README.md file I found two small errors:

1. Use this command to process an image

refers to tools/run_face_detection_on_video.py

2. Also if you are looking to conveniently process all images in one folder use this command

uses the flag --output_string which is not there. --output_folder is there though.

Great work by the way!

Hello,

I would like to know why you state "Now we have to recompile Caffe for it to work on your own machine". Why can't you do this in the Dockerfile "once and for all"?

Is there a specific Caffe related "bug" or behaviour when compiling in Dockerfile and deploying the Container on different OSes? It shouldn't be the case, as the inside OS of the Docker is always the same, right?

Kind regards,
-Christian

Why is the output video upside-down, and what are the columns contained in the video-det-fold-xxxx.txtx file ?

Hi, Can I run this in google colab? If I can please help me by providing the way to do it.

Hi, thanks for providing this! I was curious if this would with a decent frame rate on a CPU? I have a Macbook Pro with a 2.3Ghz cpu I7. Thanks very much!