If you encounter any difficulties or have any questions, feel free to ask in Issues section. I'm here to answer everyone's questions. Thank you sincerely :3.
Currently, the YOWOv3 model and paper are still being finalized, and experiments are continuously being conducted. Therefore, we are unable to provide complete official instructions at this time. However, we will provide a temporary guide, and the comprehensive official instructions will be supplemented and completed in the near future, soon :3. For now, you can:
Note: In the config folder, there are two files: ucf_config.yaml and ava_config.yaml. These are the configuration files for the corresponding datasets. We read information from these config files to build the model and specify hyperparameters and related details. To decide which file to use, simply go to utils/build_config.py and modify the default path in the build_config
function (see code below) to the desired file. This can be handled more easily using an argument parser, but as mentioned, the code is currently being used for research purposes, and everything is set up for convenience during experimentation.
def build_config(config_file='config/ucf_config.yaml'):
with open(config_file, "r") as file:
config = yaml.load(file, Loader=yaml.SafeLoader)
if config['active_checker']:
pass
return config
Currently, the displayed bounding boxes and labels are not very visually appealing. I will improve them in the near future.
python detect.py
Follow the instructions as mentioned above and run python live.py
.
Note: The Medium-1, Medium-2, and Large models correspond to the 3D backbone models shufflenetv2, i3d, and resnext101, respectively.
Some notes:
I would like to express my sincere gratitude to the following amazing repositories/codes, which were the primary sources I heavily relied on and borrowed code from during the development of this project: