tinrab / go-tensorflow-image-recognition Goto Github PK

View Code? Open in Web Editor NEW

250.0 9.0 77.0 93 KB

Image Recognition API in Go using TensorFlow

License: MIT License

Go 65.82% Dockerfile 34.18%

golang tensorflow docker api

go-tensorflow-image-recognition's Introduction

Image Recognition API in Go using TensorFlow

This is the underlying code for article Build an Image Recognition API with Go and TensorFlow.

Running the service

Build the image.

$ docker build -t localhost/recognition .

Run service in a container.

$ docker run -p 8080:8080 --rm localhost/recognition

Call the service.

$ curl localhost:8080/recognize -F 'image=@./cat.jpg'
{
  "filename": "cat.jpg",
  "labels": [
    { "label": "tabby", "probability": 0.45087516 },
    { "label": "Egyptian cat", "probability": 0.26096493 },
    { "label": "tiger cat", "probability": 0.23208225 },
    { "label": "lynx", "probability": 0.050698064 },
    { "label": "grey fox", "probability": 0.0019019963 }
  ]
}

go-tensorflow-image-recognition's People

Contributors

Stargazers

Watchers

Forkers

ardeshir tjaensch zanjs r0l1 mmkulmala chunghha stanxii gpestana vkuznet lebrave2015 adokarasev engahmed1190 gtrevg madwind76 grrrben kuailehaibin mattn happy-ferret ovace ericneid hicham1007 manuks monjovi aeneasr neopentane cbhanu arakawamoriyuki deepch markdicksonjr fmoctezuma koblas mrestivill tonicharis lanyastar athanasioskoumpouras iamvon ashneg scramble-suit multiverse-hardware-labs presleyhank geeksperiments miazzy carrot369 jiverson jiportilla lyx-outsider delfvad artdatai awesomegolang aaiyo lawhw incheol77 narayanareddy150 sml2h3 mansong1 just4ease micahli sumnotes wowplus2 rocymp jampajeen izouxv tianbingsheng isgasho kalivos icodein waldirborbajr kyong0612 litong01 ankurnitt1311 ag-go mineroot adeputraprimasuhendri eric-zsp gmgale

go-tensorflow-image-recognition's Issues

docker-compose.yaml problem

$ docker-compose -f docker-compose.yaml up -d --build
ERROR: Version in "./docker-compose.yaml" is unsupported. You might be seeing this error because you're using the wrong Compose file version. Either specify a supported version ("2.0", "2.1", "3.0", "3.1", "3.2") and place your service definitions under the `services` key, or omit the `version` key and place your service definitions at the root of the file to use version 1.
For more on the Compose file format versions, see https://docs.docker.com/compose/compose-file/

MobileNet v2 support

Hi,

Hope you are all well !

I was wondering how complicated it would be to support Mobilenet v2 - ssd with your script.

What should be modified to do object detection ?

Cheers,
X

Is there any way to recognize complex captcha?

I want to recognize captcha with golang, but I can get a solution.Do you know how to recognize captcha using go?Thank you very much!

ERROR: reference to `TF_ImportGraphDefOptionsSetDefaultDevice'

Step 14/15 : RUN go get ./... && go build -o /usr/bin/app .
---> Running in 2b689073342c

github.com/tensorflow/tensorflow/tensorflow/go

/tmp/go-build057595270/b096/_x003.o: In function _cgo_691d94bb61b7_Cfunc_TF_ImportGraphDefOptionsSetDefaultDevice': /tmp/go-build/cgo-gcc-prolog:198: undefined reference to TF_ImportGraphDefOptionsSetDefaultDevice'
collect2: error: ld returned 1 exit status
ERROR: Service 'api' failed to build: The command '/bin/sh -c go get ./... && go build -o /usr/bin/app .' returned a non-zero code: 2

Support for own trained models?

First - I love this project and thanks for sharing it.

When using my own model trained by following the tutorial here for example, I get an empty response when attempting to recognize an image - curl: (52) Empty reply from server . Any support for custom trained models?

go: error loading module requirements

Hello, I git cloned this repo and try running the example based on the instructions in the readme, however there's issues with some dependencies

$ docker-compose -f docker-compose.yaml up -d --build
Building api
Step 1/16 : FROM tensorflow/tensorflow:1.9.0
 ---> caab7ec02690
Step 2/16 : RUN curl -L    "https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-linux-x86_64-1.3.0.tar.gz" |    tar -C "/usr/local" -xz
 ---> Using cache
 ---> 159e42548844
Step 3/16 : RUN ldconfig
 ---> Using cache
 ---> acb818dd6e0d
Step 4/16 : ENV TF_CPP_MIN_LOG_LEVEL 2
 ---> Using cache
 ---> 5b5def0b688f
Step 5/16 : RUN apt-get update && apt-get install -y --no-install-recommends                g++             gcc             libc6-dev               make                pkg-config     wget     curl     git    && rm -rf /var/lib/apt/lists/*
 ---> Using cache
 ---> 40da5b287396
Step 6/16 : ENV GOLANG_VERSION 1.10.3
 ---> Using cache
 ---> 3f29629b0224
Step 7/16 : RUN set -eux;               dpkgArch="$(dpkg --print-architecture)";    case "${dpkgArch##*-}" in               amd64) goRelArch='linux-amd64'; goRelSha256='fa1b0e45d3b647c252f51f5e1204aba049cde4af177ef9f2181f43004f901035' ;;               armhf) goRelArch='linux-armv6l'; goRelSha256='d3df3fa3d153e81041af24f31a82f86a21cb7b92c1b5552fb621bad0320f06b6' ;;          arm64) goRelArch='linux-arm64'; goRelSha256='355128a05b456c9e68792143801ad18e0431510a53857f640f7b30ba92624ed2' ;;               i386) goRelArch='linux-386'; goRelSha256='3d5fe1932c904a01acb13dae07a5835bffafef38bef9e5a05450c52948ebdeb4' ;;          ppc64el) goRelArch='linux-ppc64le'; goRelSha256='f3640b2f0990a9617c937775f669ee18f10a82e424e5f87a8ce794a6407b8347' ;;               s390x) goRelArch='linux-s390x'; goRelSha256='34385f64651f82fbc11dc43bdc410c2abda237bdef87f3a430d35a508ec3ce0d' ;;               *) goRelArch='src'; goRelSha256='567b1cc66c9704d1c019c50bef946272e911ec6baf244310f87f4e678be155f2';    echo >&2; echo >&2 "warning: current architecture ($dpkgArch) does not have a corresponding Go binary release; will be building from source"; echo >&2 ;;        esac;           url="https://golang.org/dl/go${GOLANG_VERSION}.${goRelArch}.tar.gz";        wget -O go.tgz "$url";  echo "${goRelSha256} *go.tgz" | sha256sum -c -;     tar -C /usr/local -xzf go.tgz;  rm go.tgz;         if [ "$goRelArch" = 'src' ]; then                echo >&2;               echo >&2 'error: UNIMPLEMENTED';            echo >&2 'TODO install golang-any from jessie-backports for GOROOT_BOOTSTRAP (and uninstall after build)';     echo >&2;                exit 1;         fi;             export PATH="/usr/local/go/bin:$PATH";   go version
 ---> Using cache
 ---> aab9b338a518
Step 8/16 : ENV GOPATH /go
 ---> Using cache
 ---> 62940c780fc2
Step 9/16 : ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 ---> Using cache
 ---> 726bc5d09da8
Step 10/16 : RUN mkdir -p "$GOPATH/src" "$GOPATH/bin" && chmod -R 777 "$GOPATH"
 ---> Using cache
 ---> 57668e3d0dba
Step 11/16 : RUN go get -u golang.org/x/vgo
 ---> Using cache
 ---> 773c58073d44
Step 12/16 : RUN mkdir -p /model &&   wget "https://storage.googleapis.com/download.tensorflow.org/models/inception5h.zip" -O /model/inception.zip &&   unzip /model/inception.zip -d /model &&   chmod -R 777 /model
 ---> Using cache
 ---> 3a5ea3b1f9e3
Step 13/16 : WORKDIR "/go/src/github.com/tinrab/go-tensorflow-image-recognition"
 ---> Using cache
 ---> 8d79f0a4c757
Step 14/16 : COPY . .
 ---> Using cache
 ---> e07ebda3bb84
Step 15/16 : RUN vgo build -o /usr/bin/app .
 ---> Running in 5168ace03ce3
go: finding github.com/julienschmidt/httprouter v0.0.0-20180411154501-adbc77eec0d9
go: finding github.com/tensorflow/tensorflow v1.9.0
go: github.com/julienschmidt/[email protected]: unknown revision adbc77eec0d9
go: github.com/tensorflow/[email protected]: unknown revision refs/tags/v1.9.0
go: error loading module requirements
ERROR: Service 'api' failed to build: The command '/bin/sh -c vgo build -o /usr/bin/app .' returned a non-zero code: 1

Any help is appreciated

nil-operation tensorflow problem

Hi Tin!

I am trying to run your code and when I send image via curl to the server getting following issue:

http: panic serving [::1]:57414: nil-Operation. If the Output was created with a Scope object, see Scope.Err() for details.

Please note that I did not work with TF before, so I am sorry if the answer to my issue is too obvious.

lp :)

similarity search

Is it possible to return the vector before the classify layer, so that it can also be used for image similarity?

thanks

tinrab / go-tensorflow-image-recognition Goto Github PK

go-tensorflow-image-recognition's Introduction

Image Recognition API in Go using TensorFlow

Running the service

go-tensorflow-image-recognition's People

Contributors

Stargazers

Watchers

Forkers

go-tensorflow-image-recognition's Issues

github.com/tensorflow/tensorflow/tensorflow/go

Recommend Projects

Recommend Topics

Recommend Org