Sukrit Shashi Shankar

I will have to check this on the latest version of Caffe, and then perhaps get back to you !!

The loss numbers are a bit weird !! The only suggestions I can provide here are (a) please see that the labels supplied in the text file are in the...

Hi ! I am not sure of this error frankly !! I have never seen this !! I have a deadline on 24th !! I can look into this after...

This is the incorrect way of specifying labels !! I urge you to please read the README instructions to rectify this !! > On 10 Mar 2017, at 11:23, Alex...

You just have to scale the values of your bounding box parameters to a range [0,255]. Internally since we are using a data layer for labels, Caffe rescales it by...

Please check that the data you are inputting to the create_label_lmdb.py is in the format N x 20 (where N is the number of training images). Also, have you changed...

This looks correct and should not give the error that you are mentioning !! However please make sure that you are not using a single label loss like softmax but...

Also I think you should input labels in [0,255] for creating label LMDB (see README) as caffe rescales the values by 255 !! Your current label feed will not cause...

That should not affect in case you are using two data layers, one for data and one for labels as shown in the attached prototxt !!

For training I reckon phase should be TRAIN !!