#' Create the data loader
#'
#' This function loads a batch of data to be passed to the model
#'
#' @param h image height after resizing. Recommend not changing this
#' @param w image width after resizing. Recommend not changing this
#' @param file_list passed from deploy fuction
#' @param index i value from deploy function
#' @return
#'
#' @export
dataLoader <- function(file_list,
index,
w = 408, h=307){
# load image
image_path <- file_list[index]
# get image file and resize
img <- magick::image_read(image_path)
# resize image
img <- magick::image_scale(img, paste0(w, 'x', h, '!'))
# convert image to tensor. Sometimes this throws an error, so I need to catch it
img_tensor <- torchvision::transform_to_tensor(img)
# img_tensor <- tryCatch(img_tensor <- torchvision::transform_to_tensor(img),
# error = function(e) 'error')
# create a dummy target, just so we can pass something to the net
target <- torch::torch_rand(3, h, w)
# create the object that will get passed to network for this index
nn_input <- list(img_tensor, target)
return(nn_input)
}
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.