srv902 / PKU_poseattnet

Demo

To be uploaded soon

Experiments

1. Run the experiments

Just run the scripts in the bin folder.

train_dahlia_rgb.sh: Train I3D on Dahlia RGB

train_dahlia_skeletons.sh: Train LSTM on Dahlia Skeletons

train_pku_rgb.sh: Train I3D on PKU-MMD RGB

train_pku_skeletons: Train LSTM on PKU-MMD Skeletons

2. Arguments

• dataset_name: pku-mmd or dahlia
• model: i3d or lstm
• num_epochs: Number of epochs
• batch_size: Batch size for training
• eval_batch_size: Batch size for evaluation
• experiment_name: Name of the experiment
• log_dir: Directory to keep all the result in
• num_gpus: Number of GPUs

2.2 Dahlia

• dahlia_train_path: Path to the training file
• dahlia_validation_path: Path to the validation file
• dahlia_test_path: Path to the testing file
• dahlia_skeletons_dir: Directory to get the skeletos from
• dahlia_modality: rgb or skeletons

2.3 PKU-MMD

• pku_mmd_split: cross_view or cross_subject
• pku_mmd_rgb_dir: Directory to get the RGB frames from
• pku_mmd_skeletons_dir: Directory to get the skeletons from
• pku_mmd_labels_dir: Directory to get the label files from
• pku_mmd_splits_dir': Directory to get the split file from
• pku_mmd_modality: rgb or skeletons

2.4 LSTM

• dropout: Dropout
• time_steps: Number of time steps
• num_neurons: Number of neurons
• data_size: Size of the input feature vector