SymPy is not likely to accept a project under its umbrella unless it results in a lot of code for SymPy itself. I'm not saying that it's a specific rule that we have or anything, just that we tend to get a lot more students than slots, and so we have to prioritize things. Plus the PSF tends to have a ton of slots. Even so, you are welcome to direct students to submit their applications to SymPy. They will need to satisfy our application requirements same as everyone else (a merged pull request to SymPy).
If you have any interested students, go ahead and send them this way, and we can continue the discussion from there. We should probably add some addendum to our ideas page pointing to related projects' ideas pages like Theano or PyDy (feel free to go ahead and do this; it's on the wiki for a reason). Aaron Meurer On Fri, Feb 28, 2014 at 4:17 PM, Frédéric Bastien <no...@nouiz.org> wrote: > Hi, > > The current problem that we have in implementing those 2 ideas is the > time to do it. The time investment is a bit hard to justify for the people > currently developing Theano because of the high likelihood that we will > not really use those additional features for our research. > > I find those 2 ideas great and would love to see them in Theano. > I'm looking for people that could implement those features. > Our proposal to GSoC was rejected, but we got > accepted via the PSF. I don't know what will happen from this, > who will postulate, and on what they would like to work. > > Given your status as a registered GSoC organisation and the > fact that many people here would use those features, I thought > I would ask you if you were interested in helping this > get implemented. > > We can mentor/guide/plan/review all the work on those projects. > > > On a related topic, I would mention > a recent benchmark from PyDy as an example of collaboration > that we would like to encourage. That benchmark highlighted > a certain weakness in Theano that we were then able to fix > relatively quickly. > > We felt that this was very valuable and would like to see > any other such comparisions/tests performed with Theano. > This benchmark will also help with the scalar speedup project > it would serve as measurement for it. > > Thanks, > > Frédéric > > p.s. I'll be unavailable next week, so don't expect a fast answer, but > I'll be back online before the students can start with the registration > process. > > On Thu, Feb 27, 2014 at 11:42 PM, Matthew Rocklin <mrock...@gmail.com> wrote: >> I obviously think that SymPy->Theano interactions are interesting. My >> series of blogposts on that topic continue to get decent web traffic today. >> >> For what it's worth I'm also happy to vouch for Frederic and the Theano >> crew. They're a good group with which to collaborate. >> >> What sorts of interactions are you considering Fred? Are you looking for a >> co-mentor within PyDy or SymPy? Or are you just looking for commitment to >> generate problems and provide feedback? >> >> >> On Thu, Feb 27, 2014 at 7:50 PM, Jason Moore <moorepa...@gmail.com> wrote: >>> >>> They both sound like great ideas. We can certainly provide large scalar >>> functions for testing #1. I'd also like to dispatch our `solve()` methods to >>> Theano in the pydy-code-gen source. #2 is interesting, we could use >>> something like that to generate fast ODE evals that can then be wrapped >>> under a variety of high level languages (Python, Octave, R, etc). >>> >>> >>> >>> Jason >>> moorepants.info >>> +01 530-601-9791 >>> >>> >>> On Thu, Feb 27, 2014 at 9:13 PM, Frédéric Bastien <no...@nouiz.org> wrote: >>>> >>>> p...@googlegroups.com sympy <sympy@googlegroups.com> >>>> >>>> >>>> Hi, >>>> >>>> >>>> Theano is participating to GSoC via the PSF this year. In our ideas >>>> list, some of them would be very useful to PyDy/SymPy users. So I >>>> would like to bring discuss them with you. >>>> >>>> Matthew R. did a converter from SymPy graph to Theano. This allow to >>>> reuse Theano code generator. The bridge work well. Here is the 2 >>>> projects ideas that come from SymPy, PyDy users that contacted us. >>>> >>>> 1) Theano wasn't very optimized for scalar computation. But SymPy/PyDy >>>> users use very frequently scalar in computation. One of the projects >>>> idea is to lower Theano overhead. Their is a PyDy example that show >>>> well current the current overhead. I already did some optimization of >>>> that example, but their is more to do. Enough for a GSoC. >>>> >>>> 2) Generate a shared library. Many people would like to generate a >>>> shared library from a function to reuse it in other system. >>>> >>>> >>>> What do you think of that? Interested in participating in their >>>> development? >>>> >>>> Thanks. >>>> >>>> Frédéric Bastien >>>> >>>> -- >>>> You received this message because you are subscribed to the Google Groups >>>> "PyDy" group. >>>> To unsubscribe from this group and stop receiving emails from it, send an >>>> email to pydy+unsubscr...@googlegroups.com. >>>> To post to this group, send email to p...@googlegroups.com. >>>> Visit this group at http://groups.google.com/group/pydy. >>>> >>>> For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >>> -- >>> You received this message because you are subscribed to the Google Groups >>> "sympy" group. >>> To unsubscribe from this group and stop receiving emails from it, send an >>> email to sympy+unsubscr...@googlegroups.com. >>> To post to this group, send email to sympy@googlegroups.com. >>> Visit this group at http://groups.google.com/group/sympy. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/sympy/CAP7f1Ahr2crc5-qOWAg5HYr6Nyxmsb_AHu5i%3D-9%3DArU9Sx0D_g%40mail.gmail.com. >>> >>> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to sympy+unsubscr...@googlegroups.com. >> To post to this group, send email to sympy@googlegroups.com. >> Visit this group at http://groups.google.com/group/sympy. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/sympy/CAJ8oX-HZvOZRKoLh0iCCmKYOqga7RKjDKe9nDx_xBywd%3Daa3nw%40mail.gmail.com. >> >> For more options, visit https://groups.google.com/groups/opt_out. > > -- > You received this message because you are subscribed to the Google Groups > "PyDy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to pydy+unsubscr...@googlegroups.com. > To post to this group, send email to p...@googlegroups.com. > Visit this group at http://groups.google.com/group/pydy. > For more options, visit https://groups.google.com/groups/opt_out. -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to sympy+unsubscr...@googlegroups.com. To post to this group, send email to sympy@googlegroups.com. Visit this group at http://groups.google.com/group/sympy. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAKgW%3D6%2Bu975vEM7j3qqfY%3D4viwcdt-iLYgmEohqKPKFDZKvsjA%40mail.gmail.com. For more options, visit https://groups.google.com/groups/opt_out.
