models.networks.sync_batchnorm

[nvlabs]

models.networks.sync_batchnorm

[wasd96040501]

@xohvan
I think it is better to follow the README.
In second header Installation, the authors show us how to install sync_batchnorm.

Copy from README:

Clone this repo.

git clone https://github.com/NVlabs/SPADE.git
cd SPADE/

This code requires PyTorch 1.0 and python 3+. Please install dependencies by

pip install -r requirements.txt

This code also requires the Synchronized-BatchNorm-PyTorch rep.

cd models/networks/
git clone https://github.com/vacancy/Synchronized-BatchNorm-PyTorch
cp Synchronized-BatchNorm-PyTorch/sync_batchnorm . -rf
cd ../../

To reproduce the results reported in the paper, you would need an NVIDIA DGX1 machine with 8 V100 GPUs.

[claudetheboof]

wait wait .. so do you need 8 V100 gpus to test out the the code? or do you need 8 V100 gpus to train the model?

[mingyuliutw]

Train require 8 V100. Test requires 1 GPU and you need 6GB memory.

[claudetheboof]

thanks for the reply, mine was a GTX 1060 6gb ver which has just enough for the bare minimum. but does this demo included the UI like the GauGan video ?

[taesungp]

It does not contain UI like GauGAN. The code repo is designed to replicate the evaluation metrics that appear in the paper.

[Eliam76]

Hello everyone. I have exactly the same problem, using Google Colab to run SPADE. I followed the instructions from README step by step, but I still have the same issue. I am simply trying to test the training with the very limited set of images provided in the original dataset directory. Here is what I get (note that on Google Colab, OS instructions start with an exclamation mark) :

!chdir('/content/gdrive/My Drive/Colab Notebooks')
!git clone https://github.com/NVlabs/SPADE.git

Cloning into 'SPADE'...
remote: Enumerating objects: 475, done.
remote: Total 475 (delta 0), reused 0 (delta 0), pack-reused 475
Receiving objects: 100% (475/475), 21.96 MiB | 7.04 MiB/s, done.
Resolving deltas: 100% (154/154), done.
Checking out files: 100% (121/121), done.

!chdir('/content/gdrive/My Drive/Colab Notebooks/SPADE')
!cd models/networks/
!git clone https://github.com/vacancy/Synchronized-BatchNorm-PyTorch
!cp Synchronized-BatchNorm-PyTorch/sync_batchnorm . -rf
!cd ../../

Cloning into 'Synchronized-BatchNorm-PyTorch'...
remote: Enumerating objects: 16, done.
remote: Counting objects: 100% (16/16), done.
remote: Compressing objects: 100% (12/12), done.
remote: Total 177 (delta 8), reused 9 (delta 4), pack-reused 161
Receiving objects: 100% (177/177), 41.14 KiB | 249.00 KiB/s, done.
Resolving deltas: 100% (104/104), done.

!pip install -r requirements.txt

/content/gdrive/My Drive/Colab Notebooks/SPADE
Requirement already satisfied: torch>=1.0.0 in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 1)) (1.5.0+cu101)
Requirement already satisfied: torchvision in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 2)) (0.6.0+cu101)
Requirement already satisfied: dominate>=2.3.1 in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 3)) (2.5.1)
Requirement already satisfied: dill in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 4)) (0.3.1.1)
Requirement already satisfied: scikit-image in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 5)) (0.16.2)
Requirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from torch>=1.0.0->-r requirements.txt (line 1)) (0.16.0)
Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch>=1.0.0->-r requirements.txt (line 1)) (1.18.3)
Requirement already satisfied: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision->-r requirements.txt (line 2)) (7.0.0)
Requirement already satisfied: imageio>=2.3.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->-r requirements.txt (line 5)) (2.4.1)
Requirement already satisfied: scipy>=0.19.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->-r requirements.txt (line 5)) (1.4.1)
Requirement already satisfied: networkx>=2.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->-r requirements.txt (line 5)) (2.4)
Requirement already satisfied: matplotlib!=3.0.0,>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->-r requirements.txt (line 5)) (3.2.1)
Requirement already satisfied: PyWavelets>=0.4.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->-r requirements.txt (line 5)) (1.1.1)
Requirement already satisfied: decorator>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from networkx>=2.0->scikit-image->-r requirements.txt (line 5)) (4.4.2)
Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->-r requirements.txt (line 5)) (2.8.1)
Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->-r requirements.txt (line 5)) (2.4.7)
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->-r requirements.txt (line 5)) (0.10.0)
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->-r requirements.txt (line 5)) (1.2.0)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.1->matplotlib!=3.0.0,>=2.0.0->scikit-image->-r requirements.txt (line 5)) (1.12.0)

!python train.py --name 'TEST' --dataset_mode custom --label_dir './datasets/coco_stuff/train_label' -- image_dir './datasets/coco_stuff/train_img'

Traceback (most recent call last):
  File "train.py", line 12, in <module>
    from trainers.pix2pix_trainer import Pix2PixTrainer
  File "/content/gdrive/My Drive/Colab Notebooks/SPADE/trainers/pix2pix_trainer.py", line 6, in <module>
    from models.networks.sync_batchnorm import DataParallelWithCallback
  File "/content/gdrive/My Drive/Colab Notebooks/SPADE/models/networks/__init__.py", line 8, in <module>
    from models.networks.loss import *
  File "/content/gdrive/My Drive/Colab Notebooks/SPADE/models/networks/loss.py", line 9, in <module>
    from models.networks.architecture import VGG19
  File "/content/gdrive/My Drive/Colab Notebooks/SPADE/models/networks/architecture.py", line 11, in <module>
    from models.networks.normalization import SPADE
  File "/content/gdrive/My Drive/Colab Notebooks/SPADE/models/networks/normalization.py", line 10, in <module>
    from models.networks.sync_batchnorm import SynchronizedBatchNorm2d
ModuleNotFoundError: No module named 'models.networks.sync_batchnorm'

Do you have any ideas ?

[shripad1986]

ModuleNotFoundError: No module named 'models.networks.sync_batchnorm' Hi , This shows the package is missing could you search the missing package on github or from where you took other source? This could help, let me know in case of any questions Thanks again, Shripad S Kulkarni

…

On Sat, 13 Apr 2019, 05:49 xohvan, ***@***.***> wrote: i get the message ModuleNotFoundError: No module named 'models.networks.sync_batchnorm' when running test.py — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#13>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AsLPsoPWQzBBklbsfqZlLJvPlCnKgsEAks5vgSKpgaJpZM4ctvYC> .

[Eliam76]

@shripad1986 I finally found the issue and solved it. I was trying to install the package with the following code

!chdir('/content/gdrive/My Drive/Colab Notebooks/SPADE')
!cd models/networks/
!git clone https://github.com/vacancy/Synchronized-BatchNorm-PyTorch
!cp Synchronized-BatchNorm-PyTorch/sync_batchnorm . -rf
!cd ../../

unfortunately, Google Colab does not seem to handle cd command very well into the python code (and does not print the error), so the package was installed in the root directory. The following piece of code solved the problem :

running_dir = '/content/gdrive/My Drive/Colab Notebooks/SPADE'
chdir(running_dir+'/models/networks')
!git clone https://github.com/vacancy/Synchronized-BatchNorm-PyTorch
!cp Synchronized-BatchNorm-PyTorch/sync_batchnorm . -rf
chdir(running_dir)

Thank you for your help