DeepLabCut

Overview

DeepLabCut: an efficient method for 2D and 3D markerless pose estimation based on transfer learning with deep neural networks.

DeepLabCut comes in two versions CPU and GPU, each with two flavors - a command line, and a GUI version.

Currently we only have the CPU version installed. We will update this page once the GPU version has been setup.

Note: If you intend to use the GUI flavor then you can run it using the FastX service or ARCC’s new OnDemand service, Southpass.

CPU Version

Use the module name deeplabcut to discover versions available and to load the application.

Example

Module Spider Issue

If you actually spider a particular version you will see the following that indicates that you need to load a singularity version. Please note you do not need to load singularity.

[salexan5@tlog1 ~]$ module spider deeplabcut/2.2.0.2_gui -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- deeplabcut: deeplabcut/2.2.0.2_gui -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- You will need to load all module(s) on any one of the lines below before the "deeplabcut/2.2.0.2_gui" module is available to load. singularity/2.5.2 singularity/3.1.1 singularity/3.8.1 Help: DeepLabCut 2.2.0.2 GUI http://www.mackenziemathislab.org/deeplabcut/ DeepLabCut is an efficient method for 2D and 3D markerless pose estimation based on transfer learning with deep neural networks that achieves excellent results (i.e. you can match human labeling accuracy) with minimal training data (typically 50-200 frames). We demonstrate the versatility of this framework by tracking various body parts in multiple species across a broad collection of behaviors.

 

Loading and Using: DLC Light

[]$ module load deeplabcut/2.2.0.2 []$ python --version Python 3.8.12 []$ python Python 3.8.12 | packaged by conda-forge | (default, Sep 16 2021, 02:08:29) [GCC 9.4.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import deeplabcut 2021-09-28 14:27:16.934376: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /apps/s/slurm/20.11/lib64:/apps/s/slurm/20.11/lib 2021-09-28 14:27:16.934414: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. DLC loaded in light mode; you cannot use any GUI (labeling, relabeling and standalone GUI) >>> quit()

 

Loading and Using: DLC GUI

[]$ module load deeplabcut/2.2.0.2_gui []$ python -m deeplabcut 2021-09-28 14:30:07.085188: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /apps/s/slurm/20.11/lib64:/apps/s/slurm/20.11/lib 2021-09-28 14:30:07.085227: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. Starting GUI...