Installation problem (using pip in conda environment)

[aloboa]

Provide environment information

$ which python; python --version; python -m pip list | grep labelme                                
/home/alobo2/miniforge3/envs/labelmegh/bin/python
Python 3.14.2        

What OS are you using?

Ubuntu 22.04.5 LTS

Describe the Bug

As I have never used uv and use (miniforge3) conda for the rest of software, I have tried to install labelme in a conda environment.
In case using uv is actually a requirement (or even just an strong recommendation for labelme), please let me know and I will follow the uv way for labelme and this ticket can be closed (although I am concerned about eventual interference between uv and my conda environments)
This is what I do:

$ conda create -n labelmegh
$ conda activate labelmegh
$ pip install git+https://github.com/wkentaro/labelme.git

which ends with:

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
tensorflow 2.17.0 requires numpy<2.0.0,>=1.23.5; python_version <= "3.11", but you have numpy 2.2.6 which is incompatible.
Successfully installed PySocks-1.7.1 cmap-0.7.0 coloredlogs-15.0.1 filelock-3.20.3 gdown-5.2.1 humanfriendly-10.0 imgviz-2.0.0 labelme-5.11.1 loguru-0.7.3 natsort-8.4.0 onnxruntime-1.23.2 osam-0.3.1 pydantic-2.12.5 pydantic-core-2.41.5 typing-extensions-4.15.0 typing-inspection-0.4.2

$ labelme
2026-01-29 11:06:43.425 | INFO     | labelme.__main__:main:192 - Starting Labelme 5.11.1
QObject::moveToThread: Current thread (0x5e5a09778d20) is not the object's thread (0x5e5a0a230e50).
Cannot move to target thread (0x5e5a09778d20)

qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "/home/alobo2/.local/lib/python3.10/site-packages/cv2/qt/plugins" even though it was found.
This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.

Available platform plugins are: xcb, linuxfb, minimal, offscreen, vnc, webgl.

Then I try:

$ pip install pyqt
Defaulting to user installation because normal site-packages is not writeable
ERROR: Could not find a version that satisfies the requirement pyqt (from versions: none)
ERROR: No matching distribution found for pyqt
$ conda install pyqt

but get the same result.

Is there a solution to this error? Perhaps I need specific directions to install labelme in a conda environment?

Expected Behavior

labelme installed with no errors and running!

To Reproduce

No response

[wkentaro]

We officially support only installation via uv: https://labelme.io/docs/install-labelme-terminal, please use conda on your own risk

[aloboa]

ok, I confirm no problems if following directions for the uv installation (https://labelme.io/docs/install-labelme-terminal)
I was concerned with conflicts with existing conda environments.

How could I install the current github version using uv? In other words, uv equivalent to

# To install the latest version from GitHub:
# pip install git+https://github.com/wkentaro/labelme.git

(mentioned in https://github.com/wkentaro/labelme?tab=readme-ov-file#option-1-using-pip)

[wkentaro]

I think this should work:

uv tool run --python 3.13 --from git+https://github.com/wkentaro/labelme.git labelme

[aloboa]

If I do so, I get 5.11.2.dev8+g34f3aa9a8, but if, after running this command, I just run
uv tool run labelme
then I go back to Version: 5.10.1

Is this the expected/desired behavior?

[wkentaro]

I think you need --python 3.13 or set export UV_PYTHON=3.13. Otherwise uv tries to use 3.14 compatible Labelme (depending on your uv version), which the latest Labelme won't be.