EPTMainly if you need to do a lot of frontend work, is use Django.
The Big Programming Thread - Page 978
| Forum Index > General Forum |
Thread Rules 1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution. 2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20) 3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible. 4. Use [code] tags to format code blocks. | ||
|
Acrofales
Spain18331 Posts
Mainly if you need to do a lot of frontend work, is use Django. | ||
|
Manit0u
Poland17774 Posts
On November 17 2018 05:04 TentativePanda wrote: FLASK vs Django? I'm working on a university project and trying to decide which framework would be better for us (group project). Here is the info about the project and the members that are of note: - Group project (3 of us) due in 2 weeks - Need to design a user interface for our database server (website backed by DB) - MySQL was used for database design - We all have fairly decent python experience (and I'd like to think we are all capable people, willing and able to learn - even in this short amount of time) - We have little to no web development experience I know there are plenty of differences from what online (although both would work). I may be wrong, but I think the most important factor for us would be which framework allows us to work and connect to MySQL most easily. Thank you so much to people who reply, it will be super appreciated! Since you all have very little web dev experience, I strongly suggest going through this book quickly: https://www.obeythetestinggoat.com/pages/book.html#toc This will teach you all you need to know about Django (and many other things) for your project. | ||
|
LightTemplar
Ireland481 Posts
| ||
|
WarSame
Canada1950 Posts
| ||
|
Deleted User 3420
24492 Posts
I am trying to get it to work with any IDE... building with gradle. I have no idea wtf gradle is. But I can't get this to work for the life of me. I think tutorials are all out of date despite it being relatively popular. I was told to download the latest JDK. 11.0.1 But when I try to build the gradle file in intellij I just get the message "Could not determine java version from '11.0.1'." Anyone know whats going on here? edit: Well I got it working.. apparenlty updating Java messed up path variables for windows... which is really silly. and it messed them up in a pretty annoying and hard to notice way, but I eventually found it. honestly it's like oracle wants everything surrounding java to be as confusing as possible edit2: the IDEs still won't actually build a gradle project. It still says "Could not determine java version from '11.0.1'.". This is silly. | ||
|
solidbebe
Netherlands4921 Posts
| ||
|
Deleted User 3420
24492 Posts
| ||
|
WarSame
Canada1950 Posts
Also, you can build it either in Intellij(under Views -> Gradle or something like that) or from command line. | ||
|
Deleted User 3420
24492 Posts
the issue was the new JDK... java has like a 2-in-1 JDK/JRE for download all of the IDEs were not liking it | ||
|
Manit0u
Poland17774 Posts
Mercedes-Benz so crazy  | ||
|
Deleted User 3420
24492 Posts
Has it gotten to the point where people are getting hired for machine learning jobs with Bachelors degrees yet? | ||
|
Manit0u
Poland17774 Posts
On December 04 2018 22:49 travis wrote: So what languages/skills do you think would be attractive to a lot of employers in the next upcoming year in terms of employing someone out of university? Has it gotten to the point where people are getting hired for machine learning jobs with Bachelors degrees yet? People get hired for all kinds of positions with all kinds of background (half the developers I know have absolutely no CS/Engineering/Maths education whatsoever and some of them do work in ML). If you want to go for ML you'll have to know the maths behind it and being good with Python will be an asset. Here's a brief presentation by my CTO (who also works at a university and is doing research in ML). Perhaps some of the topics there will be of use to you: Edit: And some required reading: https://www.oreilly.com/library/view/introduction-to-machine/9781449369880/ https://www.oreilly.com/library/view/python-for-data/9781491957653/ https://www.oreilly.com/library/view/hands-on-machine-learning/9781491962282/ | ||
|
Deleted User 3420
24492 Posts
Python seems to be the go-to when *starting* ML, but are the businesses that are more on the cutting edge actually developing ML code in python? Isn't it slow? Or do they just not care because the slowdown doesn't matter enough to make it not worth the ease of development? | ||
|
SC-Shield
Bulgaria863 Posts
On December 05 2018 05:58 travis wrote: Side question that maybe you or someone else knows - Python seems to be the go-to when *starting* ML, but are the businesses that are more on the cutting edge actually developing ML code in python? Isn't it slow? Or do they just not care because the slowdown doesn't matter enough to make it not worth the ease of development? I don't know about businesses, but Python has some libraries written in native C, hence they're optimised. However, C/C++ should always be faster to perform, but often longer to write. It depends what you're going to write, if it's a web client, then I doubt web requests will take less time than actual performance. | ||
|
Manit0u
Poland17774 Posts
Ease of writing and changing algorithms is far more important here. Also, if you really, really care about performance you can just use cython for speeds comparable to C/C++ or move to Scala for clustering, multi-threading and scalability. I know that quite a few ML/BD projects use languages like Scala. At my company people hired for ML-related stuff usually must know Python and/or Scala (somehow other languages aren't so sought after for this field, at least where I work - my guess here would be it's because of Apache Spark). | ||
|
Neneu
Norway492 Posts
Although you can get by easily by just knowing Python, I would recommend learning R as well. A lot of people working in companies know python, but not so many know R. When it comes to statistics and machine learning R is a very powerful language, it is after all what it was created for and has a huge amount packages available and a strong scientific community. With Microsoft acquisition of R, it is also expected to be more and more easily incorporated into the .Net framework. This is the reason why I was offered the job where I work now, because I was one of the few who knows R in the area and the company (which originally was a pure web development company), is extending its strategy into fintech. When it comes to Scala I just want to say it was one of the most fun languages I've taught myself. It takes some time to learn, but is well worth it if you are going to work with Spark. However when it comes to machine learning/big data it is mostly only used with Spark, which is very specific to the type of work you are doing. You can also use Spark with R and Python, it is not until you reach a certain point you need to look under the hood of Spark and learn Scala. It should be noted though that Scala is a skill highly sought after in Finance and (strangely) Australia. So keep that in mind, if you want to pursue that. Would still recommend learning Python or R well first though. My number one recommendation would be to head over to Kaggle.com and join a competition or two. It is truly a fun experience and lets you learn a lot through said experience. It is also a great showcase for your skills as a data scientist. A few friends of mine who have created a multi million dollar company within AI in about 2 years, cares more about your Kaggle profile than your LinkedIn profile. | ||
|
Acrofales
Spain18331 Posts
And yes to all the above. Speed of your code is generally negligible in comparison to the speed of the actual algorithms, which regardless of R or Python, are heavily optimized in C. Moreover, the actual data mining part is *research*. Once you move to production, none of your code moves along, only the model, or sometimes only the results from your model will be used. It's also worth distinguishing between work/research in machine learning and data mining. Machine Learning is the development of novel algorithms, sampling methods, or ensemble techniques. This work is usually done in Matlab or C and consists mostly of improving algorithms for use in larger/different data sets, getting better accuracy, being more robust to outliers/noise/etc. Data mining is the use of these algorithms with novel data formats, and then main thrust is feature extraction and sampling, and analysis of the results. It is done mostly in R, python. Knowledge is needed of how ML algorithms work, but not how to improve them. Sometimes they overlap: eg, you're working on a DM problem and realize none of the existing ML techniques work very well, so you dig in and find a new way of doing things. Or you're working on a new ML algorithm, and find that this allows you to better solve a problem in specific types of data, so you show how well it works with novel data. But in industry, the hype right now is for data mining, not machine learning. They don't want you to invent new algorithms. They want you to apply tensorflow to their data and tell them how they can triple revenue at no cost. Regardless of that deep learning is not a suitable paradigm to apply on their Excel sheet  (can anyone tell that I'm a bitter and jaded academic) | ||
|
Silvanel
Poland4767 Posts
On December 04 2018 18:26 Manit0u wrote: https://dms.licdn.com/playback/C5600AQHlFfiT4JbEeA/d77c78bc034348d3b43524425fdaf454/feedshare-mp4_3300/1488578169071-zmy00q?e=1544004000&v=beta&t=13TPWdKrQIjvqQFDOUDzn2Z5r6ZVC-uAMFp2MoZAN6A Mercedes-Benz so crazy So i have written a post about Daimler being very good at making ads and less so about working for but it got eaten. Anyway most important thing i wanted to say that its fun and funny to drive around in test cars, You can see stickers saying: DO NOT PUSH THIS BUTTON! Max 130 km/h! Do not sit on left back seat! Airbag not working! Seatbelt not working for left backseat! Also wiring everywhere and strange devices hidden in compartments. Oh and since we have cars we work on available for employees for weekends and such - as some sort of benefit to employees its probably the same for Daimler. I know for a fact they have them in Sindelfingen. Dunno about Lisbon since it seems to be mostly backend. | ||
|
Manit0u
Poland17774 Posts
On December 05 2018 19:06 Acrofales wrote: Fuck. Microsoft bought R? There goes the neighborhood. Brb. Preparing my Linux to require .NET framework and an Azure account to run R. And yes to all the above. Speed of your code is generally negligible in comparison to the speed of the actual algorithms, which regardless of R or Python, are heavily optimized in C. Moreover, the actual data mining part is *research*. Once you move to production, none of your code moves along, only the model, or sometimes only the results from your model will be used. It's also worth distinguishing between work/research in machine learning and data mining. Machine Learning is the development of novel algorithms, sampling methods, or ensemble techniques. This work is usually done in Matlab or C and consists mostly of improving algorithms for use in larger/different data sets, getting better accuracy, being more robust to outliers/noise/etc. Data mining is the use of these algorithms with novel data formats, and then main thrust is feature extraction and sampling, and analysis of the results. It is done mostly in R, python. Knowledge is needed of how ML algorithms work, but not how to improve them. Sometimes they overlap: eg, you're working on a DM problem and realize none of the existing ML techniques work very well, so you dig in and find a new way of doing things. Or you're working on a new ML algorithm, and find that this allows you to better solve a problem in specific types of data, so you show how well it works with novel data. But in industry, the hype right now is for data mining, not machine learning. They don't want you to invent new algorithms. They want you to apply tensorflow to their data and tell them how they can triple revenue at no cost. Regardless of that deep learning is not a suitable paradigm to apply on their Excel sheet (can anyone tell that I'm a bitter and jaded academic)It is also worth noting that to pursue career in data mining you should also be pretty proficient with SQL and other query languages to fetch and aggregate data from various databases and create datasets for your algos. | ||
|
Deleted User 3420
24492 Posts
but perhaps that dream should die | ||
| ||
