BVLC/caffe

Caffe: a fast framework for deep learning. For the most recent version checkout the dev branch. For the latest stable release checkout the master branch.

Thomas A. Rieck

@trieck

What do you want to know...I’ve done a lot of on versions between Caffe and ONNX in C++. Let me know what’s on your mind.

batman-far

@batman-far

Hello,
I want to use the first five layers of Fast R-CNN plus ROI layer. And then add my new neural network to it, but I don't know how to do it. I want to pick those trained layers and then add my network.

Samjith888

@Samjith888

I'm training custom image data sets for object detection using caffe. Here i have train and validation directory. By using LabelImg , the object annotation created in a .xml file for each image separately (image1.xml,image2.xml etc.). Now i want to convert all this xml files into train.txt and test.txt. How can i do it? I stuck in this step since 3 weeks. Please help

Shreyas Moolya

@smoolya17

Can u be specific in which format do you want the bounding box to be?

Samjith888

@Samjith888

Actually I'm new to caffe. https://github.com/chuanqi305/MobileNet-SSD/issues/90#issuecomment-390118638

VOC2007
000001.jpg

The VOC2007 Database
PASCAL VOC2007
flickr
341012865

Fried Camels
Jinky the Fruit Bat

353
500
3

dog
Left
1
0

48
240
195
371

person
Left
1
0

8
12
352
498

Annotation.xml

<annotation>
<folder>train</folder>
<filename>218</filename>
<path>C:\Users\Sam\object_detection\images\train\218.jpg</path>
<source>
<database>Unknown</database>
</source>
<size>
<width>352</width>
<height>288</height>
<depth>3</depth>
</size>
<segmented>0</segmented>
<object>
<name>person</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>79</xmin>
<ymin>56</ymin>
<xmax>262</xmax>
<ymax>283</ymax>
</bndbox>
</object>
</annotation>

Neelam Shah

@NeelamShah

Hi all

Airlamb

@Airlamb

hello

getacat

@getacat

@mengpengfei 安装caffe是半年前的事了，试了很多安装方法，最后makefile运行成功，安装成功了。谢谢关心

Gloomy Ghost

@GloomyGhost-MosquitoSeal

Exastiken

@Exastiken_twitter

Hi, new to Caffe, how can I modify the individual weights of a trained caffemodel? I am planning to quantize/truncate each weight value.

Jozi Gila

@JoziGila

Hello everyone, I would like to compile this https://github.com/ShuangLI59/person_search in a jetson nano board. And I was wondering whether I could compile this library (which is the main dependency) for that board.

promach

@promach

Why ModuleNotFoundError: No module named 'caffe._caffe' ?

I already installed caffe-cpu

https://caffe.berkeleyvision.org/install_apt.html

promach

@promach

ubuntu 18.04 caffe-cpu packaged binary does not have python interface ?

promach

@promach

I have already installed python3-caffe-cpu package , but still error

https://launchpad.net/ubuntu/+source/caffe

VinodhDhan

@VinodhDhan

Can someone help me on this? https://stackoverflow.com/questions/58830041/test-net-output-0-accuracy-1-always-caffe

Always Test net output #0: accuracy = 1. My prediction is also giving same result.

Misha Lyamets

@fenchelfen

Hello, it is not super clear on docker hub how to pull the image, could you also indicate that one have to add cpu/gpu tags to pull it successfully&?

I am noting that cause I was confused at first

lenjoyn

@lenjoyn

Hi all, I am just a beginer for open source projects, and this is my first project to work on. I would like to tacke this issue: BVLC/caffe#6867 does anyone already work on it?

lenjoyn

@lenjoyn

Hi all, I think this line has something wrong:
https://github.com/BVLC/caffe/blob/04ab089db018a292ae48d51732dd6c66766b36b6/data/ilsvrc12/get_ilsvrc_aux.sh#L10

as it should be $1 instead of $0 (this is to call the name of shell)

If this is really a typo, I will make a pull request to fix it (and also make a readme.md or something to remind new users the content should be download)

avinmaulana

@avinmaulana

Hello, everyone.
It is possible to run matcaffe using newer version of matlab in ubuntu? the mex of matlab only support gcc 4.7, the gcc version of newer ubuntu only supported >4.8 (afaik).
Any ideas?
Thanks for the help

Radhika Dua

@Radhikadua123

Hi,

I could install caffe easily using conda create -n caffe_gpu -c defaults python=3.6 caffe-gpu.
But I want to compile the feature/20160617_cb_softattention branch of caffe as it contains some functions that i need to use. By following the compile steps given in the official documentation, it fails.
What steps should I follow?

Thanks in advance!

Radhika Dua

@Radhikadua123

Hi,

While installing caffe I am getting this error:

protobuf/generated_message_table_driven.h(210): error: static assertion failed with ""

1 error detected in the compilation of "/tmp/tmpxft_00007fe0_00000000-17_absval_layer.compute_61.cpp1.ii".
Makefile:585: recipe for target '.build_release/cuda/src/caffe/layers/absval_layer.o' failed
make: *** [.build_release/cuda/src/caffe/layers/absval_layer.o] Error 1

Please share how can I fix it.

Robert Nsinga

@bobuntu

Could you provide more details @Radhikadua123??

@aishangjj

Prashant Dandriyal

@PrashantDandriyal

Hello!
I am working on FeatherCNN which is based on Caffe and I can't use the library as the Caffe output blob is not accessible to me. Please help if someone has experience of FeatherCNN library...Thanks in advance

xiaofashi

@wangmaodada

Hello, the author, I have a question to consult. My former supervisor used your Caffe to make a model to identify the teapot. I didn't leave a document after leaving office. I want to learn about the use of Caffe, but the official instructions are not clear. Can you tell me how to start the identification? How do I turn off recognition?

Vladislav Denisov

@381706-1-DenisovVladislavL

Hello! Can I check whether CPU or GPU mode is supported right at runtime to set the set_mode_cpu() or set_mode_gpu() parameter respectively?

Farid

@mrgransky

Besides Caffe Model Zoo, is there other pages providing pretrained CNNs on Caffe? I found https://modelzoo.co/framework/caffe with some but details are missing !

Dhiren Hamal

@dhiren-hamal

Hello! Guys,
I want to run deeplab v2 with caffe. I have ubuntu 14.04, cuda7, cudnn4, opencv3 in docker. While compiling I got "./include/caffe/util/device_alternate.hpp:34:23: fatal error: cublas_v2.h: No such file or directory

include <cublas_v2.h>" error.

I am stuck with this problem from last a couple of days. Please help me.

Mário Almeida

@4knahs

What is the default way to run caffe models on android (arm) GPUs?

fernandoc1

@fernandoc1

In file included from /home/fccoelho/Code/caffe/src/caffe/util/math_functions.cu:1:0:
/usr/local/cuda/include/math_functions.h:54:2: warning: #warning "math_functions.h is an internal header file and must not be used directly.  This file will be removed in a future CUDA release.  Please use cuda_runtime_api.h or cuda_runtime.h instead." [-Wcpp]
 #warning "math_functions.h is an internal header file and must not be used directly.  This file will be removed in a future CUDA release.  Please use cuda_runtime_api.h or cuda_runtime.h instead."
  ^~~~~~~
In file included from /usr/local/cuda/include/cuda_runtime.h:115:0,
                 from <command-line>:0:
/usr/local/cuda/include/crt/common_functions.h:74:24: error: token ""__CUDACC_VER__ is no longer supported.  Use __CUDACC_VER_MAJOR__, __CUDACC_VER_MINOR__, and __CUDACC_VER_BUILD__ instead."" is not valid in preprocessor expressions
 #define __CUDACC_VER__ "__CUDACC_VER__ is no longer supported.  Use __CUDACC_VER_MAJOR__, __CUDACC_VER_MINOR__, and __CUDACC_VER_BUILD__ instead."
                        ^
CMake Error at cuda_compile_1_generated_math_functions.cu.o.Release.cmake:220 (message):
  Error generating
  /home/fccoelho/Code/caffe/build/src/caffe/CMakeFiles/cuda_compile_1.dir/util/./cuda_compile_1_generated_math_functions.cu.o


src/caffe/CMakeFiles/caffe.dir/build.make:517: recipe for target 'src/caffe/CMakeFiles/cuda_compile_1.dir/util/cuda_compile_1_generated_math_functions.cu.o' failed
make[2]: *** [src/caffe/CMakeFiles/cuda_compile_1.dir/util/cuda_compile_1_generated_math_functions.cu.o] Error 1
CMakeFiles/Makefile2:420: recipe for target 'src/caffe/CMakeFiles/caffe.dir/all' failed
make[1]: *** [src/caffe/CMakeFiles/caffe.dir/all] Error 2
Makefile:149: recipe for target 'all' failed
make: *** [all] Error 2

Do someone knows how to compile with CUDA 10?

I'm having that problem

yegane-AI

@yegane-AI

Guys, I wanna work with Ristretto. I've followed the manual in this link: http://lepsucd.com/ristretto-cnn-approximation/subpage/. But I couldn't find the quantized.prototxt file in the directory RistrettoDemo. Any one can help me to find that file?

Trigram19

@Trigram19

Why exactly would Caffe be a more feasible option than TF or Torch?

tahrir94

@tahrir94

Hello, I am trying to get a pre-trained caffe model for my thesis project on real-time object detection. I'm following this repo to do it: https://github.com/chuanqi305/MobileNet-SSD. So far, I annotated all the images in my datasets to generate their xml's and generated lmdb's for the test and train data. However, I just cannot find any clear instructions on how to generate a trained .caffemodel file. I would REALLY appreciate any help with this, thank you!

Ekaterina Abramova

@Katya007_gitlab

Hello, I am installing coffe using make all command, and getting error: LD -o .build_release/lib/libcaffe.so.1.0.0
ld: warning: directory not found for option '-L/Users/MYUSERNAME/anaconda/lib'
ld: warning: directory not found for option '-L/usr/local/opt/opencv4/lib'
ld: framework not found vecLib
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: * [.build_release/lib/libcaffe.so.1.0.0] Error 1
Could you please let me know how to resolve it? No help found on google.

avinmaulana

@avinmaulana

Hello, I am installing coffe using make all command, and getting error: LD -o .build_release/lib/libcaffe.so.1.0.0
ld: warning: directory not found for option '-L/Users/MYUSERNAME/anaconda/lib'
ld: warning: directory not found for option '-L/usr/local/opt/opencv4/lib'
ld: framework not found vecLib
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: * [.build_release/lib/libcaffe.so.1.0.0] Error 1
Could you please let me know how to resolve it? No help found on google.

Try using different version of opencv
make sure the directory of anaconda was set properly on you makefile.config

Ekaterina Abramova

@Katya007_gitlab

I seem to have resolved the above problem. The next issue is: ld: library not found for -lboost_python3 Any ideas on how to solve it? I can see boost_python3 folder inside my files. Where and how do I set the directory so that 'make all' stops complaining?

Kamran Tayyab

@Kamran14

https://pastebin.com/q2fWRnjk

Can someone help me out with my error?

Where communities thrive

People

Repo info

Activity

include <cublas_v2.h>" error.