hilab-git / pymic Goto Github PK

王老师，您好，在ClassificationDataset定义中，读取数据权重时是不是存在错误

if self.image_weight_idx is not None:
sample["image_weight"] = self.getweight(idx)

但是您并没有定义self.getweight。而是重写了self.get_pixel_weight
是不是应该将self.getweight(idx)修改为self.get_pixel_weight(idx

hi,你好。我已经运行了pip install pymic 和安装了requirements .txt 现在我想开始跑实例程序，就是第一个的那个蚂蚁蜜蜂，但是遇到了问题。

pymic应该是一个python包吧，为什么这里会有一个命令，叫什么pymic_train （这个命令不是linux命令）？？ 我的安装步骤是不是少了一步添加路径之类的操作啊，我以为直接pip install pymic就行了，但是好像没有添加到环境路径里面。最好是能够添加到当前的虚拟环境里面。请问要怎么添加到路径啊

When I run the Pymic_examples , I found a problem in the function logging.basicConfig in the file train.py. This function call the function datetime.now() and use it in the filename.However ,the datetime.now() function include the char ":” which is not valid in filename on windows .
I am not sure if it is the problem.
https://github.com/HiLab-git/PyMIC/blob/master/pymic/net_run/train.py
line 85

There is no config file found here, which is different from the example. How can I use this project to train my own dataset and where should the dataset be placed

A problem was encountered while running on the window system

Dear @taigw ，

Thanks for your sharing awesome projects about medical image.

Here seems should be from pymic.layer rather than from pyMIC.layer in pymic/net3d/unet3dsep.py file, line7&8, as follows
from pyMIC.layer.activation import get_acti_func
from pyMIC.layer.deconvolution import DeconvolutionLayer

make alpha dynamic

print(pymic.version) does not work

Dear author:
When I have configured the environment as required and put relevant documents in place，I run image_convert.py,but when it goes to JPCLN120.IMG or JPCLN121.IMG,it will goes wrong:
Traceback (most recent call last):
File "***/PyMIC-master/examples/JSRT/image_convert.py", line 63, in <module>
convert_JCRT_images(input_image_dir, output_image_dir)
File "***/PyMIC-master/examples/JSRT/image_convert.py", line 21, in convert_JCRT_images
data = data.reshape(shape)
ValueError: cannot reshape array of size 4130214 into shape (2048,2048)
I have no idea about this. Could you please give me a solution or suggestion？
Best wishes!

both old version and new versions have issues on colab
error in new version of pymic:
/content/drive/MyDrive/MyoPS2020/project/PyMIC/pymic/util
Traceback (most recent call last):
File "/content/drive/MyDrive/MyoPS2020/project/PyMIC/pymic/util/evaluation_seg.py", line 16, in
from pymic.util.general import is_image_name
ImportError: cannot import name 'is_image_name' from 'pymic.util.general' (/usr/local/lib/python3.10/dist-packages/pymic/util/general.py)

error in old version of pymic:(had same issue on local as well)
File "/usr/local/lib/python3.10/dist-packages/pymic/util/evaluation_seg.py", line 279, in main
evaluation(config_file)
File "/usr/local/lib/python3.10/dist-packages/pymic/util/evaluation_seg.py", line 225, in evaluation
gt_full_name = gt_root + '/' + gt_name
TypeError: can only concatenate str (not "float") to str

Dear @taigw ,

Thanks for sharing the excellent work.

It seems here should be unet3d rather than unet2d5_ag.

https://github.com/ihil/PyMIC/blob/eaeb496db257b6a58923ecdde25b25be00dd3fe4/pymic/net3d/unet2d5.py#L10

Sorry to bother.

I use pip install PYMIC for the latest 0.4.0 version.
And use it to run projects using pymic but have this issue.
together with something else, seems like I can't use

To enable the following instructions: SSE SSE2 SSE3 SSE4.1 SSE4.2 AVX AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
Traceback (most recent call last):
File "f:\ISY5001\task\MIS-FM-main\predict.py", line 10, in
from pymic.net_run.agent_seg import SegmentationAgent
File "E:\software\anaconda3\Lib\site-packages\pymic\net_run\agent_seg.py", line 20, in
from pymic.net.net_dict_seg import SegNetDict
import pymic.net.net2d.trans2d.transunet_cfg as configs
File "E:\software\anaconda3\Lib\site-packages\pymic\net\net2d\trans2d\transunet_cfg.py", line 4, in
import ml_collections
ModuleNotFoundError: No module named 'ml_collections'

How can I solve that?

nice work！
I obtained good results using a private dataset for a four-class classification [0,100,200,255]. However, I encountered a puzzling issue. When I used the command "pymic_nll test config/unet_trinet.cfg" to run the test set, the generated pred_mask values are not only [0,100,200,255], but essentially include all values from 0 to 255.
I am certain that I used four-class classification during my training and I can see the Dice coefficient for each class.

In the following lines, the config key num_workder looks like a typo of num_worker. This leads to a key error when reading the config files.

Please consider fixing it. Thanks!

nice work!
PyMIC seems to have dependencies on ml_collections, einops,monai , but these packages aren't declared in setup.py. Someone who use PyMIC may still need to run pip install ml_collections einops monai even after pip install pymic.

(pymic) ➜  ~ pip list |grep -i "pymic"
PYMIC                   0.4.0
(pymic) ➜  ~ pymic_train /home/yeep/project/py/sph/sph2d/unet.cfg
Traceback (most recent call last):
  File "/home/yeep/env/miniconda/envs/pymic/bin/pymic_train", line 5, in <module>
    from pymic.net_run.train import main
  File "/home/yeep/env/miniconda/envs/pymic/lib/python3.9/site-packages/pymic/net_run/train.py", line 10, in <module>
    from pymic.net_run.agent_seg import SegmentationAgent
  File "/home/yeep/env/miniconda/envs/pymic/lib/python3.9/site-packages/pymic/net_run/agent_seg.py", line 20, in <module>
    from pymic.net.net_dict_seg import SegNetDict
  File "/home/yeep/env/miniconda/envs/pymic/lib/python3.9/site-packages/pymic/net/net_dict_seg.py", line 24, in <module>
    from pymic.net.net2d.trans2d.transunet import TransUNet
  File "/home/yeep/env/miniconda/envs/pymic/lib/python3.9/site-packages/pymic/net/net2d/trans2d/transunet.py", line 15, in <module>
    import pymic.net.net2d.trans2d.transunet_cfg as configs
  File "/home/yeep/env/miniconda/envs/pymic/lib/python3.9/site-packages/pymic/net/net2d/trans2d/transunet_cfg.py", line 4, in <module>
    import ml_collections
ModuleNotFoundError: No module named 'ml_collections'