predict_cells_time.RdEvaluates the computational runtime of the CellFuse query prediction step.
predict_cells_time(
dataset_name,
data_dir,
test_data_dir,
test_data,
model_dir,
model_date,
device,
cluster_column,
lr,
margin,
bs,
epoch,
knn_k,
output_dim,
dropout_prob,
activation_function
)Name of the training dataset
Path to reference data directory
Path to query/test data directory
Name of the test dataset
Path to saved model directory
Date when model was created (this helps CellFuse to load correct model)
'cpu' or 'cuda'
Column name containing cluster labels
Learning rate
Margin value for contrastive loss
Batch size
Number of epochs
K for KNN classifier
Size of embedding output
Dropout probability
Activation function name
Numeric value indicating runtime (in seconds).
Useful for benchmarking CPU/GPU performance and scalability analysis.