feat: add windows env with cuda support

.gitignore

@@ -101,7 +101,6 @@ angela_merkel_speech.mp4
 shape_predictor_68_face_landmarks.dat
 
 *.mp4
-*.gif
 *.bz2
 tmp*
 *.pb

README.md

@@ -25,8 +25,12 @@ You just read original description, after this modification it is not working in
 git clone https://github.com/karolmajek/face2face-demo.git --recursive
 
 # Create the conda environment from file (Mac OSX)
-conda env create -f environment.yml
+conda env create -f environment-macosx.yml
+
+# Create the conda environment from file (Windows)
+conda env create -f environment-windows.yml
 ```
+> On Windows, make sure to switch `tensorflow-gpu` to `tensorflow` if you do not have cuda support [cuda 10](https://developer.nvidia.com/cuda-10.0-download-archive) [cuDNN v7.4.1](https://developer.nvidia.com/rdp/cudnn-archive) [instructions using cuda 9](https://medium.com/@akshaysin_86681/installing-cuda-and-cudnn-on-windows-10-f735585159f7)
 
 ### 2. Generate Training Data
 
@@ -90,7 +94,11 @@ python pix2pix.py \
   --max_epochs 200 \
   --input_dir photos/combined/train \
   --which_direction AtoB
+
+# View training progress
+tensorboard --logdir face2face-model
 ```
+> To view the training progress after starting tensorboard navigate to [http://localhost:6006/](http://localhost:6006/) 
 
 For more information around training, have a look at Christopher Hesse's [pix2pix-tensorflow](https://github.com/affinelayer/pix2pix-tensorflow) implementation.
 
@@ -123,19 +131,31 @@ For more information around training, have a look at Christopher Hesse's [pix2pi
 
     - It returns a frozen model file `frozen_model.pb` in the model folder.
 
-### 5. Run Demo
+### 5. Run Webcam Demo
 
 ```
 python run_webcam.py --source 0 --show 0 --landmark-model shape_predictor_68_face_landmarks.dat --tf-model face2face-reduced-model/frozen_model.pb
 ```
 
-Input:
-
 - `source` is the device index of the camera (default=0).
 - `show` is an option to either display the normal input (0) or the facial landmark (1) alongside the generated image (default=0).
 - `landmark-model` is the facial landmark model that is used to detect the landmarks.
 - `tf-model` is the frozen model file.
+- `save` save video to output.avi (default= 0 = false).
+
+### 5. Run Video Demo
+
+```
+python run_video.py --source path/to/video.mp4 --show 1 --landmark-model shape_predictor_68_face_landmarks.dat --tf-model face2face-reduced-model/frozen_model.pb --save 1
+```
 
+Input:
+
+- `source` path to video.
+- `show` is an option to either display the normal input (0) or the facial landmark (1) alongside the generated image (default=0).
+- `landmark-model` is the facial landmark model that is used to detect the landmarks.
+- `tf-model` is the frozen model file.
+- `save` save video to output.avi (default= 0 = false).
 
 
 ## Models
@@ -148,9 +168,15 @@ Input:
 
 Pre-trained frozen model [here](https://dl.dropboxusercontent.com/s/rzfaoeb3e2ta343/face2face_model_epoch_200.zip). This model is trained on 400 images with epoch 200.
 
+### Adrian Babilinski - Angela Merkel 1024x1024
+
+![example](example-babilinski.gif)
+
+Pre-trained updated Anglea Merkel frozen model [here](https://1drv.ms/u/s!ArYK29BEbxyk4tBpuWzwZzfiY8arRQ?e=RPzhhs). This model is trained on 400 images with epoch 200. (.98GB)
+
 ### Karol Majek - Krzysztof Gonciarz 1024x1024
 
-Me, my face and output:
+Karol Majek face and output:
 
 ![Face2Face Gonciarz](example-gonciarz.gif)
 
@@ -159,7 +185,7 @@ Video on youtube:
 
 [![Face2Face Krzysztof Gonciarz YAPZTAJ KASIĘ CICHOPEK 4 - Zapytaj Beczkę #146](http://img.youtube.com/vi/v5VDJKCrP6A/0.jpg)](http://www.youtube.com/watch?v=v5VDJKCrP6A)
 
-Frozen model can be downloaded from [here](https://goo.gl/8BgnXA) (1.1GB)
+~~Frozen model can be downloaded from [here](https://goo.gl/8BgnXA) (1.1GB)~~ _link is broken_
 
 ## Requirements
 - [Anaconda / Python 3.5](https://www.continuum.io/downloads)

environment-windows.yml

@@ -0,0 +1,51 @@
+name: face2face-demo
+channels:
+  - menpo
+  - conda-forge
+  - pytorch
+  - defaults
+dependencies:
+  - blas=1.0=mkl
+  - boost=1.63.0=py35_vc14_0
+  - bzip2=1.0.6=hfa6e2cd_1002
+  - ca-certificates=2020.4.5.1=hecc5488_0
+  - certifi=2018.8.24=py35_1001
+  - dlib=19.4=np113py35_200
+  - icc_rt=2019.0.0=h0cc432a_1
+  - intel-openmp=2019.4=245
+  - jpeg=9b=hb83a4c4_2
+  - libpng=1.6.28=vc14_1
+  - m2w64-gcc-libgfortran=5.3.0=6
+  - m2w64-gcc-libs=5.3.0=7
+  - m2w64-gcc-libs-core=5.3.0=7
+  - m2w64-gmp=6.1.0=2
+  - m2w64-libwinpthread-git=5.0.0.4634.697f757=2
+  - mkl=2017.0.4=h6d528fc_0
+  - msys2-conda-epoch=20160418=1
+  - numpy=1.13.3=py35hb69e940_3
+  - opencv3=3.1.0=py35_0
+  - openssl=1.1.1f=hfa6e2cd_0
+  - pip=9.0.1=py35_1
+  - python=3.5.3=3
+  - setuptools=27.2.0=py35_0
+  - sqlite=3.13.0=vc14_0
+  - tbb=4.4_20160413=vc14_1
+  - tk=8.5.19=vc14_0
+  - vc=14=h21ff451_11
+  - vs2015_runtime=14.16.27012=h30e32a0_2
+  - wheel=0.29.0=py35_0
+  - wincertstore=0.2=py35_1002
+  - xz=5.2.2=0
+  - zlib=1.2.8=vc14_3
+  - pip:
+    - backports-weakref==1.0rc1
+    - bleach==1.5.0
+    - html5lib==0.9999999
+    - imutils==0.4.3
+    - markdown==2.6.8
+    - protobuf==3.3.0
+    - six==1.10.0
+    - tensorflow-gpu==1.14.0
+    - werkzeug==0.12.2
+prefix: C:\ProgramData\Anaconda3\envs\face2face-demo
+