Related Paper for Segmentation
- image: (h, w, c), filter: (f_h, f_w, c), padding: P = (F - 1)/ 2, F: width of the padding, stride: S
- output_width = ((W-F) + 2 * P / S ) + 1
- Number of parameters in layers:
- Convolution layers: feature dimension: l -> k: ( m * n * l + 1 ) * k; m, n is the shape of the Convolutional filter
- Fill out the library in enviroment.yml file
- Create condata enviroments with prefix
conda env create --prefix ./env --file environment.yml # create the environment
- Activate enviroments
conda activate ./env # activate the environment
- Check package list
conda list
- Check version of the package
conda list | grep numpy
- Update package
conda env update --prefix ./env --file environment.yml --prune # update the environment
- For more information about conda
- Update torch with cuda
conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch
nvcc --version
conda activate /data.local/linhlpv/envs/aicardio
- cuda version:
nvidia-smi
- cudnn version:
/usr/include/cudnn.h | grep CUDNN_MAJOR -A 2
- cudatoolkit version:
nvcc --version