Hi, I am downloading the R packages tensorflow and keras for the first time. I am using these packages to analyze .jpg image files. Such as - How to build your own image recognition app with R! [Part 1] | R-bloggers
After installing these packages in addition to the reticulate and anaconda/miconda when I use the commands install_tensorflow(extra_packages=“pillow”), and then install_keras() after both installs, there is a Warning Message :
In readLines(rf) :
incomplete final line found on ‘C:\Users\aiucxk8\AppData\Local\Temp\RtmpsvR3Wl\file11bc4ee61bfb’
In ‘C:\Users\aiucxk8\AppData\Local\Temp\RtmpsvR3W’ folder, there is are 4 .rds files, an application file, and a folder containing the tensorflow 2.8.0 package.
In R, when I write -
setwd(“C:/Users/aiucxk8/Documents/imageDB/”)
label_list ← dir(“TRAINING/”)
output_n ← length(label_list)
save(label_list, file=“label_list.R”)
output_n
[1] 5853
When I ignore the above warning, I get an error.
path_train ← “/TRAINING/”
train_data_gen ← image_data_generator(rescale = 1/255, validation_split = .2)
2022-03-08 13:41:22.199666: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library ‘cudart64_110.dll’; dlerror: cudart64_110.dll not found
2022-03-08 13:41:22.200082: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
Loaded Tensorflow version 2.8.0
If I ignore this error, I am not able to generate the data and train on images -
width ← 50
height<- 50
target_size ← c(width, height)
train_data_gen
<keras.preprocessing.image.ImageDataGenerator>
train_images ← flow_images_from_directory(path_train,
- train_data_gen,
- subset = ‘training’,
- target_size = target_size,
- class_mode = “categorical”,
- shuffle=F,
- classes = label_list,
- seed = 2022)
Found 0 images belonging to 5853 classes.
Please help me resolve this issue - Could not load dynamic library ‘cudart64_110.dll’; dlerror: cudart64_110.dll not found
2022-03-08 13:41:22.200082: I tensorflow/stream_executor/cuda/cudart_stub.cc:29]
Thank you!