search

Neurobiologist / EC601-Pulmonary-Embolism

Development of a Deep Learning Model to Diagnose Pulmonary Embolism
3 stars 2 forks source link

Set up PENet environment #6

Closed Neurobiologist closed 4 years ago

Neurobiologist commented 4 years ago

Please install Anaconda in order to create a Python environment.

  1. Clone this repo (from the command-line: git clone git@github.com:marshuang80/PENet.git).
  2. Create the environment: conda env create -f environment.yml.
  3. Activate the environment: source activate ctpe.
Neurobiologist commented 4 years ago

After much troubleshooting, I now receive this error when creating env:

ERROR: Could not find a version that satisfies the requirement get==0.0.39 (from versions: none) ERROR: No matching distribution found for get==0.0.39

Pip subprocess error: Could not find a version that satisfies the requirement post==0.0.26

ERROR: Could not find a version that satisfies the requirement post (from versions: none) ERROR: No matching distribution found for post

Packages not available on PyPI Changed pip version in env requirements

Removed all problematic requirements. (Likely generated using pip freeze.) Kept pip version in reqs because changing it caused more errors.

Neurobiologist commented 4 years ago

Fixed path errors in source code.

Restored environment.yml and requirements.txt. Re-create conda env. Update: Errors still exist. See below.

Neurobiologist commented 4 years ago

Pip subprocess error: Could not find a version that satisfies the requirement get==0.0.39

Neurobiologist commented 4 years ago

Restoring modified env and reqs

Fixed issues with argparse and test.sh script

Neurobiologist commented 4 years ago

THCudaCheck FAIL file=torch/csrc/cuda/Module.cpp line=32 error=10 : invalid device ordinal Traceback (most recent call last): File "test.py", line 144, in args_ = parser.parse_args() File "/projectnb2/ece601/kaggle-pulmonary-embolism/meganmp/PENet/args/base_arg_parser.py", line 115, in parse_args torch.cuda.set_device(args.gpu_ids[0]) File "/usr3/graduate/meganmp/.conda/envs/ctpe/lib/python3.6/site-packages/torch/cuda/init.py", line 262, in set_device torch._C._cuda_setDevice(device) RuntimeError: cuda runtime error (10) : invalid device ordinal at torch/csrc/cuda/Module.cpp:32

Fixed GPU issue. test.sh --gpu_ids 0

Neurobiologist commented 4 years ago

File "test.py", line 145, in test(args_) File "test.py", line 29, in test data_loader = CTDataLoader(args, phase=args.phase, is_training=False) File "/projectnb2/ece601/kaggle-pulmonary-embolism/meganmp/PENet/data_loader/ct_data_loader.py", line 16, in init dataset = dataset_fn(args, phase, is_training) File "/projectnb2/ece601/kaggle-pulmonary-embolism/meganmp/PENet/datasets/ct_pe_dataset_3d.py", line 47, in init with open(args.pkl_path, 'rb') as pkl_file: FileNotFoundError: [Errno 2] No such file or directory: '/projectnb/ece601/kaggle-pulmonary-embolism/meganmp/PENet/data/series_list.pkl'

[Troubleshooting] This pkl file is not being generated anywhere.

Neurobiologist commented 4 years ago

Emailed the corresponding author of the paper to request access to the pkl files. [Pending response] Update 10/15: Corresponding author responded and cc'd someone who should be able to help. Update 10/16: Ongoing correspondence to resolve issue.

Neurobiologist commented 4 years ago

Ongoing correspondence to resolve problem. Raising separate issue.

Environment set up successfully. Closing issue #6.