Are you sure you want to delete this access key?
Experiments are done on window laptop with GTX 1060.
Duplicate my environment with conda
using environment config (.yml
files)
conda env create -f <config-filename>.yml
base-env.yml
: bare bone environment that can use GPU computationfull-env.yml
: full environment used in this projectIf you wanna install everything yourself. Follow instructions below
For GPU computation and basic needs
Our environment will be named 'kangaroo'
Python: 3.8
conda create -n kangaroo python=3.8
Activate: conda activate kangaroo
Check python version
python --version
> Python 3.8.15
⭐ Tensorflow-GPU:
conda install tensorflow-gpu==2.5.0
(2.3.0) does not work !! at all This install also included numpy (and keras)
(Extra) Install cuda-nvcc
(Reference errors and fixes)
conda install -c nvidia cuda-nvcc
(Optional) Check and sync/downgrade tensorflow-base
version
To match with tensorflow version
conda install tensorflow-base==2.5.0
Verify installation packages recognizes your GPU
python src/scripts/verify_gpu.py
Continue install all other packages
Clone submodule and install Mask-RCNN
# clone
git submodule init
git submodule update
# install
cd src/submodules/akTwelve_MaskRCNN
python setup.py install
Exception:
Mask_RCNN
: LinkOther used packages:
conda install -c conda-forge [packages]
Pillow
: Image loading/processing (also the base of Keras
Image)scikit-image
: (! version: 0.16.2
)ipykernel
pycocotools
: Needed for training on windows (pip
)dvc
imgaug
: image augmentation frame work used by MRCNN training methodimageio
: helper image reading libpandas
: data frame manipulation (test related purposes)Error: When installing pycocotools
: Missing C++ stuffs
Pitfall: Wrong installation of build tools / installation with mismatching C++/SDK version / missing components
Error: module 'keras.engine' has no attribute 'Layer'
Keras conflict with scikit-image
in RCNN code
Input image dtype is bool. Interpolation is not defined with bool data type
scikit-image
from 0.19.3
to 0.16.2
conda install ... --force-reinstall
pip install -U ...
Getting stuck with Error No module named 'tensorflow_core.estimator'
Tensorflow 2.1 conflict with h5py
Warining: Internal: Failed to launch ptxas
cuda-nvcc
Call to CreateProcess failed. Error code: 2
1:
Press p or to see the previous file or, n or to see the next file
Are you sure you want to delete this access key?
Are you sure you want to delete this access key?
Are you sure you want to delete this access key?
Are you sure you want to delete this access key?