Still having drawing bug where masks wont actually be created once drawn

[mayishazn]

I created a new environment and did a fresh install. The red circle does disappear once I close the loop but it doesn't turn into a mask. I've drawn four here and they're just stuck with the purple dotted lines. There is no drawing related output in the terminal either.

Originally posted by @mayishazn in #961 (comment)

Also sorry for the repeat issue. It would just be highly convenient if I could fix masks on cellpose 3 directly.

[carsen-stringer]

oh no, I was hoping I had found the bug! Does this happen when you start from scratch: can you delete your _seg.npy file and start labeling again and confirm it happens? can you also please export your environment conda env export > cp.yml and share it here? I will try to reproduce. how big is your image?

[mayishazn]

name: freshcellpose3
channels:

• pytorch
• nvidia
• defaults
  dependencies:
• blas=1.0=mkl
• bzip2=1.0.8=h2bbff1b_6
• ca-certificates=2024.9.24=haa95532_0
• console_shortcut=0.1.1=haa95532_6
• cuda=11.6.1=0
• cuda-cccl=11.6.55=0
• cuda-command-line-tools=11.6.2=0
• cuda-compiler=11.6.2=0
• cuda-cudart=11.6.55=0
• cuda-cudart-dev=11.6.55=0
• cuda-cuobjdump=11.6.124=0
• cuda-cupti=11.6.124=0
• cuda-cuxxfilt=11.6.124=0
• cuda-libraries=11.6.1=0
• cuda-libraries-dev=11.6.1=0
• cuda-memcheck=11.8.86=0
• cuda-nsight-compute=12.4.1=0
• cuda-nvcc=11.6.124=0
• cuda-nvdisasm=12.6.77=0
• cuda-nvml-dev=11.6.55=0
• cuda-nvprof=12.6.80=0
• cuda-nvprune=11.6.124=0
• cuda-nvrtc=11.6.124=0
• cuda-nvrtc-dev=11.6.124=0
• cuda-nvtx=11.6.124=0
• cuda-nvvp=12.6.80=0
• cuda-runtime=11.6.1=0
• cuda-sanitizer-api=12.6.77=0
• cuda-toolkit=11.6.1=0
• cuda-tools=11.6.1=0
• cuda-version=12.6=3
• cuda-visual-tools=11.6.1=0
• expat=2.6.3=h5da7b33_0
• fontconfig=2.14.1=hb33846d_3
• freetype=2.12.1=ha860e81_0
• gmpy2=2.1.2=py311h7f96b67_0
• intel-openmp=2023.1.0=h59b6b97_46320
• jinja2=3.1.4=py311haa95532_0
• libcublas=11.9.2.110=0
• libcublas-dev=11.9.2.110=0
• libcufft=10.7.1.112=0
• libcufft-dev=10.7.1.112=0
• libcurand=10.3.7.77=0
• libcurand-dev=10.3.7.77=0
• libcusolver=11.3.4.124=0
• libcusolver-dev=11.3.4.124=0
• libcusparse=11.7.2.124=0
• libcusparse-dev=11.7.2.124=0
• libffi=3.4.4=hd77b12b_1
• libglib=2.78.4=ha17d25a_0
• libiconv=1.16=h2bbff1b_3
• libnpp=11.6.3.124=0
• libnpp-dev=11.6.3.124=0
• libnvjpeg=11.6.2.124=0
• libnvjpeg-dev=11.6.2.124=0
• libpng=1.6.39=h8cc25b3_0
• libuv=1.48.0=h827c3e9_0
• libxml2=2.13.1=h24da03e_2
• mkl=2023.1.0=h6b88ed4_46358
• mpc=1.1.0=h7edee0f_1
• mpfr=4.0.2=h62dcd97_1
• mpir=3.0.0=hec2e145_1
• mpmath=1.3.0=py311haa95532_0
• nsight-compute=2024.3.2.3=0
• openssl=3.0.15=h827c3e9_0
• pcre2=10.42=h0ff8eda_1
• pip=24.2=py311haa95532_0
• python=3.11.10=h4607a30_0
• pytorch-cuda=11.6=h867d48c_1
• pytorch-mutex=1.0=cpu
• pyyaml=6.0.2=py311h827c3e9_0
• setuptools=75.1.0=py311haa95532_0
• sqlite=3.45.3=h2bbff1b_0
• tbb=2021.8.0=h59b6b97_0
• tk=8.6.14=h0416ee5_0
• typing_extensions=4.11.0=py311haa95532_0
• tzdata=2024b=h04d1e81_0
• vc=14.40=h2eaa2aa_1
• vs2015_runtime=14.40.33807=h98bb1dd_1
• wheel=0.44.0=py311haa95532_0
• xz=5.4.6=h8cc25b3_1
• yaml=0.2.5=he774522_0
• zlib=1.2.13=h8cc25b3_1
• pip:
  • cellpose==3.1.1.dev1+g9398cac
  • colorama==0.4.6
  • fastremap==1.15.0
  • filelock==3.16.1
  • fsspec==2024.10.0
  • imagecodecs==2024.9.22
  • llvmlite==0.43.0
  • markupsafe==3.0.2
  • natsort==8.4.0
  • networkx==3.4.2
  • numba==0.60.0
  • numpy==2.0.2
  • opencv-python-headless==4.10.0.84
  • packaging==24.1
  • pillow==10.2.0
  • pygments==2.18.0
  • pyqt6==6.7.1
  • pyqt6-qt6==6.7.3
  • pyqt6-sip==13.8.0
  • pyqtgraph==0.13.7
  • qtpy==2.4.2
  • roifile==2024.9.15
  • scipy==1.14.1
  • superqt==0.6.7
  • sympy==1.13.1
  • tifffile==2024.9.20
  • torch==2.5.1+cu118
  • torchaudio==2.5.1+cu118
  • torchvision==0.20.1+cu118
  • tqdm==4.66.6
  • typing-extensions==4.12.2
    prefix: C:\Users\mayis\anaconda3\envs\freshcellpose3

[mayishazn]

These are 2048x2048 images and some of them are zstacks with 10 slices.
It seems to happen regardless of if I'm starting from scratch or fixing masks that cyto3 found.

[mayishazn]

idk if the image type matters? These are tiffs

[carsen-stringer]

For the stacks are you opening them with --Zstack? In that case single stroke is off

[mayishazn]

Yes I am opening them with --Zstack. So how do I edit them then?

[carsen-stringer]

Ahha that's it, you have to press enter to complete the mask, or you can turn on single stroke