Autotune | Automatic calibration methodology

I am following TFF tutorials to build my FL model My data is contained in different CSV files which are considered as different clients. Following this tutorial, and build the Keras model function as following

...

I am reproducing the examples of the chapter 16 of the book Hands-On Machine Learning of Aurélien Géron and found an error while trying to train a simple RNN model.

The error is the following:

...

What I'm trying to achieve is to simulate a streaming learning method using Tensorflow's fit() and evaluate() methods.

What I have until now is a script like this, after getting some help from the community here:

...

I have a dataset made of tensors. A sample tensor looks like this:

...

I am new in federated learning I am currently experimenting with a model by following the official TFF documentation. But I am stuck with an issue and hope I find some explanation here.

I am using my own dataset, the data are distributed in multiple files, each file is a single client (as I am planning to structure the model). and the dependant and independent variables have been defined.

Now, my question is how can I split the data into training and testing sets in each client(file) in federated learning? like what we -normally- do in the centralized ML models The following code is what I have implemented so far: note my code is inspired by the official documentation and this post which is almost similar to my application, but it aims to split the clients as training and testing clients itself while my aim is to split the data inside these clients.

...

I am training a Unet segmentation model for binary class. The dataset is loaded in tensorflow data pipeline. The images are in (512, 512, 3) shape, masks are in (512, 512, 1) shape. The model expects the input in (512, 512, 3) shape. But I am getting the following error. Input 0 of layer "model" is incompatible with the layer: expected shape=(None, 512, 512, 3), found shape=(512, 512, 3)

Here are the images in metadata dataframe.

Randomly sampling the indices to select the training and validation set

...

Using Tensorflow's Dataset generator without repeat works. However when I use repeat to double my train dataset from 82,000 to 164,000 for additional augmentation I "run out of data."

I've read that steps_per_epoch can "slow cook" models by allowing multiple epochs for a single pass through training data. It's not my intent, but even when I pass a small number of steps_per_epoch (which should create this slow cooking pattern), TF says I've ran out of data.

There is a case where TF says I'm close ("in this case, 120 batches"). I've attempted plus/minus this value but still getting errors with drop_remainder set to True to drop anything left over.

Error:

WARNING:tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least steps_per_epoch * epochs batches (in this case, 82,000 batches). You may need to use the repeat() function when building your dataset. WARNING:tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least steps_per_epoch * epochs batches (in this case, 120 batches). You may need to use the repeat() function when building your dataset.

Parameters Train Dataset 82,000 Val Dataset 12,000 Test Dataset 12,000 epochs (early stopping usually stops about 30) 100 batch_size 200

**batch_size is the same for model mini-batch and generator batches

Attempt steps_per_epoch Value Error steps_per_epoch==None None "..in this case, 82,000 batches" steps_per_epoch==train_len//batch_size 820 "..in this case, 82,000 batches" steps_per_epoch==(train_len//batch_size)-1 819 Training stops halfway "..in this case, 81,900 batches" steps_per_epoch==(train_len//batch_size)+1 821 Training stops halfway "..in this case, 82,100 batches" steps_per_epoch==(train_len//batch_size)//2 410 Training seems complete but errors before validation "..in this case, 120 batches" steps_per_epoch==((train_len//batch_size)//2)-1 409 Same as above:Training seems complete but errors before validation "..in this case, 120 batches" steps_per_epoch==((train_len//batch_size)//2)+1 411 Training seems complete but errors before validation "..in this case, 41,100 batches" steps_per_epoch==(train_len//batch_size)*2 1640 Training stops at one quarter "..in this case, 164,000 batches" steps_per_epoch==20 (arbitrarily small number) 20 Very surprisingly "..in this case, 120 batches"

Generators - goal is to repeat the train set two times:

...

I'm trying to train a neural network made with the Keras Functional API with one of the default TFDS Datasets, but I keep getting dataset related errors.

The idea is doing a model for object detection, but for the first draft I was trying to do just plain image classification (img, label). The input would be (256x256x3) images. The input layer is as follows:

...

I am trying to mimic the federated learning implementation provided here: Working with tff's clientData in order to understand the code clearly. I reached to this point where I need clarification in.

...