Are there any inference code for Inception resnet v2 object detection model

[yt605155624]

Are there any inference code for Inception resnet v2 object detection model ?just input the model and a picture and get a picture with boxes and classes,thank you !

[rkrasin]

@yt605155624 can you point to the model you have in mind?

[rkrasin]

Most likely, you will be served by https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/oid_inference_and_evaluation.md

[yt605155624]

@rkrasin Sorry,I have seen this before,but it said I have to pack the val/test data into TFRecords and finally get a mAP,what I want to do is to test a picture,where Input a picture and a pretrained model,then I get a detecetd picture with boxes classes and the scores(0~1.0),can you help me ?

[rkrasin]

@yt605155624 ah! I would recommend to use this Jupyter Notebook as the reference:
Object Detection Demo. That notebook can run on individual images for any model from the Model Zoo, including two models trained on OpenImages.

[yt605155624]

@rkrasin Thank you very much ,By the way ,can you tell me how to Download Images With Bounding Boxes Annotations ,I have followed the link https://github.com/cvdfoundation/open-images-dataset ,after install gsutil(I have got the access from http://www.cvdfoundation.org/datasets/open-images-dataset/signup.html) ,I login my gmail with the command "gcloud auth login [yt605155624] --no-launch-browser",then ,I got a link ,and entered the verification code ,then,ERROR HAPPENED, timed out!!!! I am in China,but I have climb over the wall，I want to know ,are My COMMAND ERROR ,OR JUST CANNOT DOWNLOAD FROM GOOGLE-CLOUD FROM CHINA? BY THE WAY,are there any other way to download the datasets,for example ,just a link for downloading ,thank you!

[rkrasin]

@yt605155624 to be honest, you might have a better luck asking on relevant Chinese forums. Most importantly, because they would better know the quirks related to the Great Chinese Firewall.

Last time I visited China, I had hit a few similar issues here and there and resorted to using sshuttle. Not sure, if it would be of any help for you.