v-sundaresan/truenet

DL tool for white matter hyperintensities segmentation

truenet's Issues

Training the TrUE-Net model from scratch but fail to converge

Hello,
I used the truenet train command to train, but the loss of the verification set did not decrease, and the dice value remained unchanged.
Do you have any suggestions?

Problem accessing 'lit' library

I have encountered the following issue on multiple computers (Ubuntu 23.04, Python3).

It seems that there are problems interacting with 'lit'.

However my environment reports the following upon 'pip install lit':
Requirement already satisfied: lit in /home/chris/venv/lib/python3.11/site-packages (16.0.6)

Do you have any suggestions for overcoming it?

------------------------------------------------------------------ERROR MESSAGE-----------------------------------------------------------------

truenet evaluate
Traceback (most recent call last):
File "/usr/local/bin/truenet", line 4, in
import('pkg_resources').run_script('truenet==1.0.1', 'truenet')
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 3248, in
@_call_aside
^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 3223, in _call_aside
f(*args, **kwargs)
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 3261, in _initialize_master_working_set
working_set = WorkingSet._build_master()
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 617, in _build_master
ws.require(requires)
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 956, in require
needed = self.resolve(parse_requirements(requirements))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 815, in resolve
dist = self._resolve_dist(
^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 856, in _resolve_dist
raise DistributionNotFound(req, requirers)
pkg_resources.DistributionNotFound: The 'lit' distribution was not found and is required by triton

Gpu Recognition

Hello,
I used Anaconda Navigator to setup an env with Python 3.6.13 and torch 1.5.0 and get the following error. Notably, CUDA is True and the GPU is available.
Advice appreciated.
During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/usr/local/bin/truenet", line 4, in
import('pkg_resources').run_script('truenet==1.0.1', 'truenet')
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 708, in run_script
self.require(requires)[0].run_script(script_name, ns)
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 1528, in run_script
exec(script_code, namespace, namespace)
File "/usr/local/lib/python3.11/dist-packages/truenet-1.0.1-py3.11.egg/EGG-INFO/scripts/truenet", line 156, in
File "/usr/local/lib/python3.11/dist-packages/truenet-1.0.1-py3.11.egg/truenet/true_net/truenet_commands.py", line 277, in evaluate
File "/usr/local/lib/python3.11/dist-packages/truenet-1.0.1-py3.11.egg/truenet/true_net/truenet_test_function.py", line 98, in main
File "/usr/local/lib/python3.11/dist-packages/truenet-1.0.1-py3.11.egg/truenet/utils/truenet_utils.py", line 65, in loading_model
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 795, in load
return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 1012, in _legacy_load
result = unpickler.load()
^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 958, in persistent_load
wrap_storage=restore_location(obj, location),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 215, in default_restore_location
result = fn(storage, location)
^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 182, in _cuda_deserialize
device = validate_cuda_device(location)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/torch/serialization.py", line 166, in validate_cuda_device
raise RuntimeError('Attempting to deserialize object on a CUDA '
RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.
(truenet) roger@roger:~$ python3.6
Python 3.6.13 |Anaconda, Inc.| (default, Jun 4 2021, 14:25:59)
[GCC 7.5.0] on linux
Type "help", "copyright", "credits" or "license" for more information.

import torch
print("CUDA available: ", torch.cuda.is_available())
CUDA available: True
print("Number of GPUs available: ", torch.cuda.device_count())
Number of GPUs available: 1

TypeError: can only concatenate str (not "torch.device") to str

Hi, thanks for this this great tool! We followed the basic installation instructions and would like to use one of the trained models, but we keep getting a strange concatenation error. It looks like device isn't a string, so in places where you have print('testfunction:device used:' + device) maybe print('testfunction:device used:' + str(device)) would work?

Our command and the output are pasted below.

truenet evaluate -i testing-truenet-dvs/sub-10866_testing -m mwsc -o testing-truenet-dvs/sub-10866_results
FLAIR image found for sub-10866
Traceback (most recent call last):
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/bin/truenet", line 4, in <module>
    __import__('pkg_resources').run_script('truenet==1.0.1', 'truenet')
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/lib/python3.11/site-packages/pkg_resources/__init__.py", line 720, in run_script
    self.require(requires)[0].run_script(script_name, ns)
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/lib/python3.11/site-packages/pkg_resources/__init__.py", line 1570, in run_script
    exec(script_code, namespace, namespace)
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/lib/python3.11/site-packages/truenet-1.0.1-py3.11.egg/EGG-INFO/scripts/truenet", line 159, in <module>
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/lib/python3.11/site-packages/truenet-1.0.1-py3.11.egg/truenet/true_net/truenet_commands.py", line 409, in evaluate
  File "/ZPOOL/data/tools/anaconda/tug87422/anaconda3/lib/python3.11/site-packages/truenet-1.0.1-py3.11.egg/truenet/true_net/truenet_test_function.py", line 38, in main
TypeError: can only concatenate str (not "torch.device") to str

Docker

We are very interested in applying this to our acute stroke data sample (over 3000 participants).
Are there any plans for a docker image?
I have been having trouble getting both the preparation and evaluation commands to recognize the GPU and proper CUDA libraries (leading to 2 separate Ubuntu reinstalls!0
Thanks!
Roger