资源算法deepdreamer

deepdreamer

2019-12-30 | |  35 |   0 |   0

Deep Dreamer

Easy to configure Python program that make use of Google's DeepDream

Requirements

Installation

  1. Install NumPy, SciPy, Pillow and Caffe. On Ubuntu 17.10 installing caffe will usually install all other dependencies.

  2. Clone this project. git clone https://github.com/kesara/deepdreamer.git

  3. Go to project directory. cd deepdeamer

  4. Download deploy.prototxt from bvlc_googlenet into the project directory.

  5. Add line force_backward: true to deploy.prototxt file.

  6. Download bvlc_googlenet.caffemodel from bvlc_googlenet into the project directory.

  7. (Optional) Download MIT's "Places" neural net, download the Places205-GoogLeNet from their website. You need the deploy_places205.protxt and googlelet_places205_train_iter_2400000.caffemodel files from the archive.

Usage

  • Just deep dreamingpython3 deepdreamer.py image.jpg

  • Create a deepdream gifpython3 deepdreamer.py --gif true image.jpg

  • Create a deepdream video (requires ffmpeg)python3 deepdreamer.py --video video.mp4

Configuration options

usage: deepdreamer.py [-h] [--zoom {true,false}] [--scale SCALE]
                      [--dreams DREAMS] [--itern ITERN] [--octaves OCTAVES]
                      [--octave-scale OCTAVE_SCALE] [--layers LAYERS]
                      [--clip {true,false}] [--gpuid GPUID]
                      [--network {bvlc_googlenet,googlenet_place205}]
                      [--gif {true,false}] [--reverse {true,false}]
                      [--duration DURATION] [--loop {true,false}]
                      [--framerate FRAMERATE] [--list-layers] [--video VIDEO]
                      [image]

positional arguments:
  image

optional arguments:
  -h, --help            show this help message and exit
  --gpuid GPUID         enable GPU with id GPUID (default: disabled)
  --zoom {true,false}   zoom dreams (default: true)
  --scale SCALE         scale coefficient for zoom (default: 0.05)
  --dreams DREAMS       number of images (default: 100)
  --itern ITERN         dream iterations (default: 10)
  --octaves OCTAVES     dream octaves (default: 4)
  --octave-scale OCTAVE_SCALE
                        dream octave scale (default: 1.4)
  --layers LAYERS       dream layers (default: inception_4c/output)
  --clip {true,false}   clip dreams (default: true)
  --network {bvlc_googlenet,googlenet_place205}
                        choose the network to use (default: bvlc_googlenet)
  --gif {true,false}    make a gif (default: false)
  --reverse {true,false}
                        make a reverse gif (default: false)
  --duration DURATION   gif frame duration in seconds (default: 0.1)
  --loop {true,false}   enable gif loop (default: false)
  --framerate FRAMERATE
                        framerate for video (default: 24)
  --list-layers         list layers
  --video VIDEO         video file

Examples

DeepdreamDeepdreamDeepdreamDeepdream


上一篇:DeepDreamVideo

下一篇:deepdream-docker

用户评价
全部评价

热门资源

  • seetafaceJNI

    项目介绍 基于中科院seetaface2进行封装的JAVA...

  • spark-corenlp

    This package wraps Stanford CoreNLP annotators ...

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • shih-styletransfer

    shih-styletransfer Code from Style Transfer ...