Caffe Users

I am finding this error in the testing part. The training part went well. For the same net and same

1. Yes, this is exactly how it works! 2. No difference at all from the mathematical point of view. It

Hello, Can you please help with an issue I am facing while using caffe . I want to keep data in

I am experiencing a similar problem. What I have seen so far is that first image gets stuck in the

argmax arguments are the number of returns you want argmax function to return. for example, if i

I'm using a custom data layer provided in a fork to read video frames. This implementation reads

the input data is n*c*7*7 blob，now I want to simultaneously use some crops of the inputdata wrt

W, H and finally C, for a total of N times, that's it El lunes, 14 de noviembre de 2016, 15:01:33

I suppose your data is of the wrong shape. Caffe expects a 4D blob at the input, made of (batch_size)

Hi, I am a very new user to Caffe. As part of my research I am doing an experiment to modify a NN to

The `Silence` layer type takes bottoms without making any tops to ignore unattached tops like the `

Just in case someone is interested in this, I finally found out what I was doing wrong. I was really

Hello, I have been trying to get the output of the network using C++. It contains a 40 channel

Hey, Thank you so much for sharing. I still have some confusion about setting the delta value of 0 or

Just figured it out. In case anyone comes across similar issues, here's the solution: The issue

Each layer in caffe will call forward_gpu. By default forward_gpu calls forward_cpu as it is expected

Do you want a blob that is of size 501x64? If that is the case you should reshape your scale factor

Thank you Jan, ad 1) OK, actually It not so important in my task. ad 2) Sure, I understand that I

Thank you - this is great news! Being able to use more training examples should greatly improve my

The update is performed several stages, first in the solver (eg SGDSolver.cpp) in SGDSolver<Dtype

Thank you very much for the detailed answer! Very helpful! Den torsdag 29 oktober 2015 kl. 11:05:45

On reducing the vocabulary size (number of output classes) to 82994, training continued smoothly. But

It's weird that sometimes the mean of label blob is right, sometimes it's wrong. What's

Thank you! Interesting to know. I also found this paper: https://arxiv.org/abs/1511.07376 so I know

Hi All- Two quick questions: - What is the best way to instantiate a new blob Y for a loss layer,

Did this work? On Wednesday, 21 October 2015 16:56:15 UTC+5:30, Jan wrote: Pickling is only supported

Arghavan, here is a sample code for working with matlab wrapper: addpath(genpath('matlab'));

Hi, I have a network which looks like this: data:blob->layer1:layer->blob1:blob->........-

Not exactly. It totally does make sense to run the loops this way because that's how data is

Hi, I am experiencing with the same error message. Did you find the solution? Thanks. On Thursday,