<center>![Resources #17.png](https://steemitimages.com/DQmQp2V1Ug7GVfcMKL8bGD2D6C3skLW4nYkx4FksPNJZURS/Resources%20%2317.png)</center>
___
Adam Bouhenguel has recently put up a library that you can use in Python for deep learning. It's generically called _tfi_ and I assume it's a combination of the words 'TensorFlow' and 'Inception'. 

As per Adam:

_"TFI provides a simple Python interface to any TensorFlow model. It does this by automatically generating a Python class on the fly."_ [[source](https://github.com/ajbouh/tfi)]

In the example provided in his github repository he uses TFI with a saved model of [Inceptionv1](http://nicolovaligi.com/history-inception-deep-learning-architecture.html), which is a popular convolutional neural network used in a host of deep learning projects. 

I wanted to create a video demonstration of the example provided by Adam, but since he doesn't provide a download link to the weights (the pretrained model), it might not be easy to replicate and it might take some time to train the model in the exact way in which he trained. 

I think this might be a barrier to the popularity of the repo, unless he's going to make a priority of this matter. I asked him about this on reddit but it's been 2 days and i got no reply so far. If he replies, I'd probably do a demo quickly. 

Anyway, the code, once you have the saved model, is really simple, and makes deep learning child's play. If you're interested, you can check out the repository following the link below:
___
<center>[Using TensorFlow Models in One Line of Code](https://github.com/ajbouh/tfi)</center>
___
### <center>To stay in touch with me, follow @cristi </center>  
___
[Cristi Vlad](http://cristivlad.com), Self-Experimenter and Author