You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Now here's a really fascinating one. Weekly number of research papers indexed on arXiv specifically, and on all of Google Scholar (which includes arXiv), that mention TensorFlow, Keras, or PyTorch.pic.twitter.com/oVBsukcqXg
This tells us two things: 1) The pure ML research community (posts on arXiv) is evenly split between TF and PyTorch 2) The broader scientific community interested in applying ML (doesn't post on arXiv; often doesn't even publish preprints), which is ~20x larger, uses TF/Keras
The former leads the latter, however. As the pure ML research community commits one way or the other *over the long term*, the applied ML in other fields will eventually follow out of necessity. Strong network and incumbency effects at play here.
I'm not so sure. Given that published results can usually be replicated in either framework, but software engineering and infrastructure usually have stronger requirements, I think the research community doesn't have that much say on what's used in production.
Can confirm the statement 2 for big scale medical workflow, the major reason we chose tf/keras is because of its docs and overall compatibility.
I am happy to see these results but I would refrain from including SO questions. Why? Because it doesn't differentiate between:
1. Unable to understand the api usage?
2. v1-> v2 migration problem?
3. Something confusing/buggy?
Saying it again: We should have official forums 
Every data source is biased in some way. Which is why we look at many sources to see the real picture
I use TF/keras because the pipeline to getting it on mobile devices is easier. The keras packaging of layers also lets me play network surgery in a way that is less easy than on pytorch.
This is pretty interesting and it’s going in the counter direction of what people are thinking (the rise of pytorch). I don’t know if it’s possible but getting some insights from github repositories might reveal more information about this.
You might need to exclude papers that mentioned both (tensorflow and pytorch) and ended up using one. Unless you can figure out which one is used in the paper and not just mentioned 
as literature review!
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
