Home

Awesome

Repo-2021

<b>Anaconda</b>
Here you can find how to successfully install Anaconda, Google Cloud libraries and OpenCV in Ubuntu 20, as well as drivers for a GPU NVIDIA RTX 2060

<b>Arduino</b>
In Arduino folder, there is a bash command to solve permissions issues in Arduino, as well as the code for a Temperature and Humidity sensor.

<b>Autorun</b>
Here there are commands for creating a service in Ubuntu.

<b>BOVESPA</b>
Here there is a code that analyzes Fibonnaci, RSI and closing price for stocks in Brazilian Stock Exchange, saving engineered variables to BigQuqery in order to create a dashboard in Data Studio.

<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/BOVESPA/bovespa.png>

<b>C-Language</b>
Notebooks of C language of my course on Udemy. There is also a Cellular Automata code in C that is intended to be used in LEDs.

<b>Cellular Automata - One-Dimensional Cellular Automata from Scratch </b>
This part has codes to develop Elementary Cellular Automata in Python, binary and 5 states, and also an experiment of adding a CA to a PyTorch CNN's kernel, to solve the MNIST task. The model achieves 99.29 accuracy <a href="url">https://github.com/RubensZimbres/Repo-2021/tree/main/Cellular_Automata/Test/Learning/Tese/99.29t</a>

<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/Cellular_Automata/CA1D_5_.png>

<b>Google-Cloud-Deploy</b>
Here are the commands to successfully deploy a model in Google Cloud AI Platform.

<b>Google-Cloud-Functions-Job</b>
Here there are commands and Python files to deploy in Google Cloud Functions.

<b>Google-Cloud-IAP</b>
Here are the directions to securely connect your Jupyter Notebook to a Google Cloud Compute Engine via internal IP through an IAP (Identity-Aware Proxy) tunnel + TCP Forwarding. Full tutorial here: <a href="url">https://www.linkedin.com/pulse/connecting-jupyter-notebook-google-cloud-vm-using-only-rubens-zimbres/</a>

<b>Google-Cloud-OCR</b>
Code for submitting documents to Document AI, OCR service from Google Cloud.

<b>Google-Cloud-Storage</b>
Python code to rename all files in storage, even if they have strange characters, as 'ç' and 'õ'.

<b>Google-Cloud-Train</b>
Instructions to submit a training job for AI Platform and inference.

<b>Convolutional Graph Neural Networks in PyTorch </b>
This folder brings Graph CNNs in PyTorch with dgl library. There is an Attention Graph Neural Net, as well as Graph Neural Nets for Classification and to find communities.

<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/Graph-Networks/graph4_comm_movie.gif>

<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/Graph-Networks/movie.gif>

<b>Hugging Face</b>
Here there are codes to query GPT-2, for the Squeeze BERT (a smaller version of BERT).

<b>Kali</b>
Here there is a bunch of bash and Python scripts for ethical hacking, development of exploits and exploiting OWASP Top 10.

<b>Keras</b>
Codes for LSTM and simple neural network in Keras Sequential.

<b>Neural_Cellular_Automata_Google</b>
Here there is the code used by Google guys to create a Neural Cellular Automata, using Sobel filter.

<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/Neural_Cellular_Automata_Google/output_batches.gif>

<b>OpenCV</b>
A simple algorithm in OpenCV to detect filled circles in school tests:
<img src=https://github.com/RubensZimbres/Repo-2021/blob/main/OpenCV/works.png>

<b>OpenCV_Haarcascade</b>
Here are the instructions to train your own object detector using haarcascade in OpenCV.

<b>Prophet</b>
Code to make predicions with Facebook's Prophet.

<b>PySpark</b>
A sample notebook in PySpark with some basic codes in example.py and ML classifiers .

<b>PyTorch Geometric</b>
Graph Neural Networks in PyTorch Geometric, for protein and molecule structure.

<b>Python Vulnerabilities</b>
This part shows how to install Bandit, a library that scans Python codes for common vulnerabilities to malicious hackers.

<b>RRG</b>
This code shows how to prepare data for RRG - Relative Rotation Graph, for stocks analysis of momentum and relative strength.

<b>Tensorflow</b>
Here there are codes of Gated Recurrent Units for Text Generation in Keras and links for the hierarchical Softmax activation function.

<b>Transformer</b>
In this section you will find codes to customize PyTorch architecture, load partial weights, a transformer implementation, beam search, generation of word embeddings and also a Transformer customized to act as a time series predictor, using floats instead og long integers, as the original.

<b>ONNX</b>
A tool to speed up inference time of PyTorch models.

<b>Root dir: file Embeddings_PyT_TF.py</b>
How to generate word embeddings (3,768) in PyTorch and Tensorflow.

<b>Root dir: file GPU_GCP.txt</b>
How to install NVIDIA driver in Google Cloud Compute Engine.