Sports-Recognition
Please visit the projects section for tasks.
The Website - Sports Recognition
Collection of sports images covering 73 different sports.. Images are 224,224,3 jpg format. Data is separated into train, test and valid directories. Additional a csv file is includes for those that wish to use it to create there own train, test and validation datasets.
Model Summary:
Model: "model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_layer (InputLayer) [(None, 224, 224, 3)] 0
_________________________________________________________________
data_augmentation_layer (Seq (None, None, None, 3) 0
_________________________________________________________________
efficientnetb0 (Functional) (None, None, None, 1280) 4049571
_________________________________________________________________
global_average_pooling (Glob (None, 1280) 0
_________________________________________________________________
output_layer (Dense) (None, 73) 93513
=================================================================
Total params: 4,143,084
Trainable params: 874,953
Non-trainable params: 3,268,131