gru | Gru : automatic management of Docker containers | Continuous Deployment library

I am trying to use Powershell to replace a semicolon ; with a pipe | that is in a file that is semicolon separated, so it's a specific set of semicolons that occurs between double-quotes ". Here's a sample of the file with the specific portion in bold:

Camp;Brazil;AI;BCS GRU;;MIL-32011257;172-43333640;;"1975995;1972871;1975";FAC0088/21;3;20.000;24.8;25.000;.149;GLASSES SPARE PARTS,;EXW;C;.00;EUR;

I've tried using -replace, as follows:

I know you can use different types of layers in an RNN architecture in Keras, depending on the type of problem you have. What I'm referring to is for example layers.SimpleRNN, layers.LSTM or layers.GRU.

So let's say we have (with the functional API in Keras):

I've been trying to speed up training of my CRNN network for optical character recognition, but I can't get the accuracy metric working when using TFRecords and tf.data.Dataset pipelines. I previously used a Keras Sequence and had it working. Here is a complete runnable toy example showing my problem (tested with Tensorflow 2.4.1):

I am trying to get the second last value in each row of a data frame, meaning the first job a person has had. (Job1_latest is the most recent job and people had a different number of jobs in the past and I want to get the first one). I managed to get the last value per row with the code below:

first_job <- function(x) tail(x[!is.na(x)], 1)

first_job <- apply(data, 1, first_job)

I use my custom dataset class to convert audio files to mel-Spectrogram images. the shape will be padded to (128,1024). I have 10 classes. after a while of training in the first epoch, my network will be crashed inside the hidden layer in GRU shapes due to this error:

I am trying to create an autoencoder that is capable of finding anomalies in text sequences:

I am new to deep learning and I try to implement an RNN (with 2 GRU layers). At first, the network seems to do it's job quite fine. However, I am currently trying to understand the loss and accuracy curve. I attached the pictures below. The dark-blue line is the training set and the cyan line is the validation set. After 50 epochs the validation loss increases. My assumption is that this indicates overfitting. However, I am unsure why the validation mean absolute error still decreases. Do you maybe got an idea?

One idea I had in mind was that this could be caused by some big outliers in my dataset. Thus I already tried to clean it up. I also tried to scale it properly. I also added a few dropout layers for further regularization (rate=0.2). However these are just normal dropout layers because cudnn does not seem to support recurrent_dropout from tensorflow.

Remark: I am using the negative log-likelihood as loss function and a tensorflow probability distribution as the output dense layer.

Any hints what I should investigate? Thanks in advance

Edit: I also attached the non-probabilistic plot as recommended in the comment. Seems like here the mean-absolute-error behaves normal (does not improve all the time).

I am running tensorflow 2.4 on colab. I tried to save the model using tf.train.Checkpoint() since it includes model subclassing, but after restoration I saw It didn't restored any weights of my model.

Here are few snippets:

I am currently working on a encoder-decoder model using GRUs. It takes 2 inputs, encoder input and decoder input. There is only one output from the decoder. The model is: