environment-detector.yml

# Use this environment to run `detection/run_tf_detector*.py`
# If your computer does not have a CUDA-supported GPU, change the
# dependency `tensorflow-gpu` to `tensorflow`

name: cameratraps-detector

dependencies:
- python>=3.5
- nb_conda_kernels
- ipykernel
# note that tensorflow < 1.13.0 needs CUDA 9.0, and later versions need CUDA 10;
# depending on the CUDA stack installed in your environment, you might have to specify
# a particular version of tensorflow in the range given here:
- tensorflow-gpu>=1.9.0, <1.15.0
- tqdm
- jsonpickle
- Pillow>=5.3.0
- humanfriendly
- numpy
- matplotlib
- nb_conda_kernels
- ipykernel
- opencv

# the `nb_conda_kernels` and `ipykernel` packages are installed so that we
# can use Jupyter Notebooks with this environment as a kernel