EdgarAndresSantamaria / Deep_Learning_Introduction-Tensorflow-

Deep_Learning_Introduction-Tensorflow-

This repository aims to provide some state-of-art Deep Learning approaches implemented from scratch, those notebooks have explanations to aid users finishing up their own approaches, first we should start generating a Google Drive account and install Colaboratory complement, in order to connect your Gogle Drive to your Colaboratory environment, you need to do the following at the main Google Driv page:

• Click "New" and look for the option "Connect more apps".
• Click "Connect more apps" and look for "Colaboratory". Click "+ CONNECT", and follow instructions.

Then we provide a blog to introductory get used to it https://towardsdatascience.com/getting-started-with-google-colab-f2fff97f594c. After setting up the required enviroment, each user should download this repository and upload into their rspective Google Drive accounts, it's important to understand previous steps for further steps all over the laboratories, and some usual problems are related with paths https://medium.com/lean-in-women-in-tech-india/google-colab-the-beginners-guide-5ad3b417dfa.

Now, we are able to start with some baby steps over 'labs/' (easy tasks) folder, this should be easy following https://colah.github.io/posts/2015-08-Understanding-LSTMs/. Finally we provide an attention lab (intermedium difficulty) that should be understanded following Rocktäschel's.

Finally we provide some practical approaches using the begginers previous information with some data scientist steps over 'assigments/' (intemedium tasks) folder, here, every user will be able to apreciate the powerfulness of the provided tools. And there are also provide implementations of many usual Deep Learning models as RNNs or Word Embbedings. The last model provided 'word-by-word-attention' (hard tasks) will ensemble toguether all concepts with some informatics engineering steps and should be understanded reading all the pappers provided inside.

Note: the general aim of this contribution is to aid further research and also play with the models (maybe applying optimization strategies ...), for any issue try to contact [e-mail]

Thanks all and hope to contribute Deep Learning development, cheers.

Data access -> https://drive.google.com/drive/folders/1YrzY2BtNvNVBW2cooYt32r1hAOaG2Plc?usp=sharing

Team information:

-labs -> Edgar Andrés

-assigments -> Edgar Andrés, Mohammed Yassin, Xaidé Caceres and Radostina Peteva

special thanks to Oier Lopez de Lacalle and Olatz Perez de Viñaspre.