TensorFlow 2.0 upgrade, Python support, & more!

Can we take a look at the first one that came in was from alpha Arthur Arthur. Can I ask about any pre-built binary for the RTX 2080 GPU on Ubuntu 16 specs. So I like specific question time to be associated with a specific driver from Nvidia. So like the version of true that we support of the version of cudnn that we support. So my recommendation would be if you're taking a look at any of the pre-built fineries take a look at what driver or what version of the driver you have supported on that specific card. I'm

not an expert in Nvidia card so that I love them. So I don't really know what supported by that card Arthur but if you go over here like on them on my laptop, I have I called up like some of the what Nvidia say as their tensorflow system requirements and the specific versions of the drivers that they support the one got you and we had The last second is wild that I find when working with gpus is that it's easy for you to go to the driver vendor and download the latest version, but that may not be the one that time flow is built for the 1 to 10. So just make sure that they actually match each

other and you should be good to go even with that particular card warm feelings and thoughts about builds in general for tensorflow. We have a great special interest groups specifically focused on that called Sig build build. So strongly suggest going to the community section of our GitHub in checking out the the Sig Bill Bliss service start of joining it in in joining our our weekly stand-ups. So the next question is a really funny when I think that how many times have you been asked this today at least 12:00

bullies. The other flavor of it is well is this particular symbol that I use all the time is this going to also be supported? And if not, what is changed invested so much time building stuff in tents for 1. Axe. They don't want to do I find scrubs work with tensorflow to probably not they they would not work with with tensorflow 2.0 out of the box, but we have created an upgrade utility for you to use its automatically downloaded with tensorflow too. Oh whenever you download it for more

information on it and what in particular it's doing you can check out this medium blog posts that I and my colleagues Anna created as well as this upgrade to tensorflow 2.0 video goes through and with gifts which is the best communication shows you how you can use the upgrade script on an infant. So any sort of arbitrary python file or even jupyter notebooks one of our machine learning GDs created in Tension that allow you to do that as well and it'll give you an expert. Text file on that shows you all of the symbol renames the added

keywords and then also some manual changes if you have to make manual changes, usually you do not sow to see this in action we can go and we can go and take a look at this particular text generation example that we have running a Shakespeare wall takes all of the Corpus of Shakespeare text trains against the Shakespeare text and generates something that The Bard could have potentially written, you know, should he have had access to to to deep learning app to deep learning resources on the Oaks you were off your

idol 3. Video about nothing you could exports the python file and then to upgrade it looks like the requirements of are you at we can check to see if it were using tensorflow Alpha and then like I mentioned before all you would have to do is practice this with a pointy if I praise be to the name of the the name of the Python file was text generation. I want to create an upgrade shift enter. It does its it does it's upgrading magic and very quickly and tells me all of the things that we need to be changed to make it too. Oh compatible and creates that file for me

off to the side. So now if I wanted to if I wanted to run this model, it should be able to It should be able to train as it as it what so let's just check to make sure that would be the case, but I'm breaking changes within the AP. You can see that you have some renames and some additional key words. Are there any gifts for those of us who don't say GIF when it comes to upgrade? There are few little. Use basically in summary, but hopefully this blog post and your video and all the stuff that we doing will

help you get around those doctors before we've had such an interest in testing tensorflow too. Oh and trying it out against historic model that we've formed a weekly testing stand up. And also we have a migration. Support our that's being on the thing implemented with the internal support our so if you have an external group to Google that's interested in upgrading your models, please join the testing group and and we can get you situated on a lot of stuff that we've seen like an example, while she was training fashioned

fudge. It's like there's a lot of the surface level code that you're going to be riding in Carousel PlayStation change it move on to the next question that we can talk all day, but we just mention carrots and it appears. So what is the purpose of keeping estimators in Charis as separate apis? Is there going to be something native Securus models that allows a distributor training Allied training evaluate? The many purposes right? So I think for me the main purpose that I would like to think of though is one that is because a lot of people are using them

as quickly as so it's like an inside when it comes to like estimator that's tomatoes are really great for large-scale painting a lot of times when I first started I started like this cuz I couldn't figure out what an old wasn't a neural network and other all these Concepts that I had to learn a simple estimator that I could use to the to do like it at the end of something like that. So, you know, they're there for a reason to lie down on it has to go to and I think like we just spoke about

that the code is the same between one and two and it's at the layers AP I think makes it super simple for you to design a neural network on in the fact that you can go low. Levels beyond that like you don't get to find your own. There's really give you that allows you to drive stick instead of driving automatic. And then if you need to do additional customizations, there's a subclass in component. And then if you need to go even lower then we have something called TF module and we even expose some

of the basic almost core Ops tensorflow as well. They're really it at any sort of level you want to interact with the API you can but there is like that there's something called distributed strategy and the whole idea behind that allow you to be able to distribute your training maybe across multiple gpus on the same machine maybe across multiple gpus on different machines, maybe across CPU spread all over the place. Caritas we love them both that both still there. Hopefully, this is something that will help you with a question. I think we got time for just one

more. So this is a page question. I am the python person so ask tensorflow when will tensorflow be supported in Python 3 7 and hence be expressed in Anaconda 3 so I can certainly answer the Python 3 7 and also I would love to speak a little bit more about support for python going forward to answer the question. I'm going to bounce over to our tensorflow 2.0 project tracker. These are all of the standing issues that we have when doing development for tensorflow too. It's transparency

and all of them are transparent to the public. So if you ever want to have a sort of context on where we stand currently what we have yet to do this project tracker is a great way to Go to understand that but let's take a look at three 7 and there we go. So then process of releasing binaries for Python 3 5 and 3 7 does issued 25 for 20 and it's going a little bit off the screen for 29, but you can you can take a look at that issue and see that was currently in progress. There's not really an ETA but it's something

that we want to have complete by the time that the Elsa RC is really so so that is wonderful to see there's also a website called python3 statement statement. Com. Maybe it's. Org. Tensorflow is made the commitment that as of January 1st 2020. We no longer support python to and we have done that with a plethora of our of our python community. So tensorflow pandas scikit-learn Etc. We are firmly committed to python3 and Python 3 support. So so you will you will be getting your Python 3 support and we are we are firmly committed to having the

most thing about the issue tracker was going to be a big heavy case of it's totally transformed. You can keep an eye on what we're doing, Sounds good. Okay, I think that's all we have time for us. Whatever you do. Don't forget to hit that subscribe button or I am. Thank you so much and thanks for being engaged.