Image-Classification | short template to classify images | Machine Learning library

I have a DataFrame like this:

I am trying to recreate CNN image recognition model from this paper(model 1) using different images. However, fitting the model returns me a ResourceExhaustedError at the first epoch. The batch size is already considerably small so I'm guessing the problem is with my model definition which I've copied from the paper. Any advice on what to change with the model will be appreciated. Thank you!

I try to make a page to upload some files to server side. But it does not work.

I am trying to model for the first time in GCP and I cannot find nor figure out how I link the data to model on. In my script, I would typically write read_csv from this path.

I get that I have to load it to Google Cloud Storage. Its a csv and I'm running xgb classification over it. The question is how to link those things so the script knowss to run it on that...

I am making a neural network architecture to predict whether an image is of a building, forest, glacier, mountain, sea or street. Link to the dataset.

I am using tf.keras.preprocessing.image.ImageDataGenerator() to load and preprocess my data. The directory look like this:

Dataset Directory

The directory contains images of size 150x150 like this:

Photo from class 'Building'

I am following this tutorial with my custom data and my custom S3 buckets where train and validation data are. I am getting the following error:

I am learning the implementation of the TensorFlow model at android. In this tutorial, it put the labels.txt and model.tflite files into assets folder . https://blog.notyouraveragedev.in/android/image-classification-in-android-using-tensor-flow/ What is that labels.txt should be? I have a file that has the following format :

Background:

I'm building a real time image classification app for food items with react-native and the tensorflow HOC Camera.

I'm using this tutorial as a basis which includes a github repo where you can test it via Expo:

Basically, I want the camera preview to remain live and display predictions from the tensorflow model live. There is a function that is called each frame and predicts it. I can console.log() the predictions live but displaying them using a state variable causes the camera to re-render every prediction.

This is obviously poor user experience as only the text containing the prediction should change, the camera doesn't need to rerender. I've tried using React.memo() but that provides a 'Performance hint' and does not guarantee the camera component will not rerender.

Code:

I'm following the "Getting started with the built-in image classification algorithm" tutorial from Google's AI Platform and before submit a training job, one has to specify (it seems optional) the "image_size" which is defined as: "the image size (width and height) used for training". Do I have to specify a couple of scalars (comma delimited?) or a scalar ? How can we interpret this parameter? If I specify something, does it impose that all my input images should be of this particular size or will the images be automatically resized (or cropped?) to this size by the training graph? And equivalently for prediction task do I have to resize my input image to this specific size or does the prediction graph takes care of that.