Comments (83)
Hi everyone,
I have some good news!!
I have been working on the Ruby API for Tensorflow and I have some interesting findings that I would like to share. The tensorflow.rb gem can be found on somaticio/tensorflow.rb.
Aside from that, I have written three blog posts where I have given a very brief summary of the work
The project still needs a lot of work and contributions are very welcome. I encourage everyone to read the blog posts and then install and play with Tensorflow.rb. Any comments/suggestions would be nice and you can always post your thoughts on the gitter channel.
Special Thanks to Jason Toy (Somatic.io), Soon hin khor and Sameer Deshmukh for mentoring and supporting me throughout the project.
from tensorflow.
Tensorflow Ruby API is one of the projects submitted to GSoC 2016 by
SciRuby. It is undergoing the selection process, and the results will be
out soon (22/23 Apr).
On Sun, Apr 17, 2016 at 2:18 AM, Riley Usagi [email protected]
wrote:
What about now?
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
Hello,
Jason (@jtoy on GitHub) has volunteered to sponsor a student to write a Ruby TensorFlow wrapper over 3 months this summer. SciRuby will be providing technical assistance in this endeavor and
People interested in winning this grant should go over this blog entry for further details: http://www.somatic.io/blog/tensorflow-is-coming-to-ruby
Applicants are expected to provide a detailed proposal about the approach they will employ for writing the Ruby wrappers and provide monthly timeline over the next 3 months starting 1st of June. Deliverables should be on a monthly basis. We expect a weekly time line and proper evaluation of the approaches that you think are the best for this project.
Proposals will be accepted till 27 May. Please make sure you send us your proposals before this date.
The sponsorship amount will be paid in monthly installments subject to completion of work for that particular month. The payment will be delayed if the applicant cannot deliver the promised material.
We also looking for additional sponsors for this project so interested people can get in touch by mailing me at [email protected].
Do spread the word!
from tensorflow.
I will pay for a Ruby api.
2015-11-12 11:39 GMT+01:00 Babak Gh [email protected]:
[image: π]
β
Reply to this email directly or view it on GitHub
#50 (comment)
.
Manuel Arno Korfmann
Luetzowstrasse 8
24105 Kiel
manu @ korfmann . info
+49 40 210913788
from tensorflow.
I don't think that was meant to be an exhaustive list...
from tensorflow.
Hi Everyone,
@nethsix informed me about this opportunity.
I think it would be great to develop this project since so many people are interested in this project.
I am sharing my GSoC project proposal please give me your valuable suggestions.
from tensorflow.
+1
from tensorflow.
+1. I am willing to work on this, it would be good to coordinate with anyone else interested.
from tensorflow.
Hi @jtoy and all potential sponsors:
I'm Sameer from SciRuby. I am this year's GSOC admin. After some discussion we have decided that we would be happy to provide any technical assistance to you and the student should you decide to take this project up.
I will send you a detailed email soon elaborating on how we can proceed forward with this issue, since a GitHub issue is really not the best place to discuss something like this.
In the meantime, please join the SciRuby mailing list to get a heads up on the current progress of projects and our work methodology in general.
from tensorflow.
+1 I'd be interested in contributing, too! :)
from tensorflow.
What about now?
from tensorflow.
Hey Akshay,
GSoC requires students to specify the deliverables each week for the entire
GSoC period, e.g., which API will be wrapped, created, etc. This enables
mentors to track progress accurately, and ensure students stick to their
commitments. I am not sure if it is permissible to share the plan wholesale
but if you think about it, this basically involves looking at Tensorflow
docs, tutorials, etc. to see what is the minimum amount of API to get
started, e.g., model a simple linear regression y = W.x + b requires
tf.placeholder, tf.Variable, tf.matmul, etc. and work from there. The
applicant has also created a repo to wrap the version API, i.e., return the
Tensorflow version, which may seem simple, but it acts a good foundation to
build from. Let me ask the student to share it, and maybe SciRuby can fork
it so that everyone can contribute from there. How does that sound? Any
suggestions?
khor
On Mon, May 2, 2016 at 3:15 PM, Akshay Khole [email protected]
wrote:
Hi @nethsix https://github.com/nethsix can we take a look at the
detailed plan? Is it possible to make it available to us?β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
Tensorflow Ruby API is very important for a large number of people from around the world (for example I'm from Russia), so maybe a good option would be to start a sort of open-raise funds for participants GSoC who wanted to undertake the development of this project, but so that the source was discovered during development, so that everyone can make a contribution. (sorry for my bad English)
from tensorflow.
+1
from tensorflow.
Guys, thanks a lot for working on this; is it possible to keep us updated on the progress, here or anywhere else? π
from tensorflow.
+1
from tensorflow.
As you have rightfully guessed, the core TensorFlow team doesn't have plans for Ruby on the horizon, so if you want to contribute it, please go ahead! Submitting drafts for frontends such as this on the discuss mailing list early on is probably a good idea.
from tensorflow.
+1 Let me know of any efforts. I will for sure be forking that.
from tensorflow.
hi all, im very serious about getting this ported to ruby. I do not have
the expertise to do that though. I would be willing to fund part (maybe
all) of this if someone or some team wants to work on it. Are any of you
interested in working on this? I believe besides the SWIG bindings
integration, we need the numpy like interface for manipulating the data.
My reasons for wanting this is simple. ruby has been cutoff from all the
scientific computing, but rubyists are great innovators and have great
ideas. Tensorflow is a new software paradigm that is going to level the
playing field in deep learning and bring about a lot of new and wonderful
innovations. I want ruby community to be able to participate. Please feel
free to email me,(check my profile) if interested in this.
On Fri, Mar 4, 2016 at 10:00 AM, Pavel Suchmann [email protected]
wrote:
Great idea.
Some clever numpy equivalent should be used for TF/Ruby with respect to
interfacing TF tensors.
https://github.com/SciRuby/nmatrix has an interface resembling numpy...?β
Reply to this email directly or view it on GitHub
#50 (comment)
.
from tensorflow.
@JeffBusterCase Great! We can learn from one another and move this forward.
from tensorflow.
Cool! I hope it would be accepted in oder that Ruby users can much enjoy
deep learning with the simple interfaced and powerful Tensorflow package. I
am now using it everyday for researching like medical drug discovery.
Especially I am using Tensorflow with Skflow which is renamed recently as
'learning' underneath of Tensorflow umbrella. So, I hope Ruby-Tensoflow may
include Skflow interface as well. What do you guys think?
On Sat, Apr 16, 2016 at 9:50 PM Neth Six [email protected] wrote:
Tensorflow Ruby API is one of the projects submitted to GSoC 2016 by
SciRuby. It is undergoing the selection process, and the results will be
out soon (22/23 Apr).On Sun, Apr 17, 2016 at 2:18 AM, Riley Usagi [email protected]
wrote:What about now?
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<
https://github.com/tensorflow/tensorflow/issues/50#issuecomment-210860945>β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#50 (comment)
Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
from tensorflow.
π
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
+12
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
I += 1
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
Has anyone written a swig interface yet for tensorflow? I would like to investigate writing this.
from tensorflow.
+1
from tensorflow.
+1
from tensorflow.
How's going on this subject?
Regards,
James
[email protected]
Best regards,
Sungjin (James) Kim, Ph.D.
CCB in Harvard
[email protected]
On Mon, Jan 25, 2016 at 8:31 PM, Ahn, Soohan [email protected]
wrote:
- 1
β
Reply to this email directly or view it on GitHub
#50 (comment)
.
from tensorflow.
+1. At your service.
from tensorflow.
Interested, but is there a technical reason it needs to be SWIG? I can work in raw Ruby/C, Ruby/C++ fine, but find SWIG a bit too dense.
from tensorflow.
How about using RICE?
On Mon, Feb 1, 2016 at 5:12 PM Neil Slater [email protected] wrote:
Interested, but is there a technical reason it needs to be SWIG? I can
work in raw Ruby/C, Ruby/C++ fine, but find SWIG a bit too dense.β
Reply to this email directly or view it on GitHub
#50 (comment)
.
from tensorflow.
+1
from tensorflow.
Great idea.
Some clever numpy equivalent should be used for TF/Ruby with respect to interfacing TF tensors.
https://github.com/SciRuby/nmatrix has an interface resembling numpy...?
from tensorflow.
hi jtoy
I think NMatirx and SciRuby can be alternative tools for numpy. What do you
think?
On Fri, Mar 4, 2016 at 10:49 AM jtoy [email protected] wrote:
hi all, im very serious about getting this ported to ruby. I do not have
the expertise to do that though. I would be willing to fund part (maybe
all) of this if someone or some team wants to work on it. Are any of you
interested in working on this? I believe besides the SWIG bindings
integration, we need the numpy like interface for manipulating the data.
My reasons for wanting this is simple. ruby has been cutoff from all the
scientific computing, but rubyists are great innovators and have great
ideas. Tensorflow is a new software paradigm that is going to level the
playing field in deep learning and bring about a lot of new and wonderful
innovations. I want ruby community to be able to participate. Please feel
free to email me,(check my profile) if interested in this.On Fri, Mar 4, 2016 at 10:00 AM, Pavel Suchmann [email protected]
wrote:Great idea.
Some clever numpy equivalent should be used for TF/Ruby with respect to
interfacing TF tensors.
https://github.com/SciRuby/nmatrix has an interface resembling numpy...?β
Reply to this email directly or view it on GitHub
<
https://github.com/tensorflow/tensorflow/issues/50#issuecomment-192311660>
.β
Reply to this email directly or view it on GitHub
#50 (comment)
.
Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
from tensorflow.
I've seen both of those libraries and I know they have been around for a few years. I've never personally used them, but I know they are supposed to be alternatives to numpy/scipy. If we feel they can meet our needs, then I advocate that we use them instead of rewriting everything ourselves.
from tensorflow.
This is going to happen in Google Summer of Code 2016 (https://github.com/SciRuby/sciruby/wiki/Google-Summer-of-Code-2016-Ideas), or/and probably earlier (not sure if GSoC requires students to be on board before starting).
from tensorflow.
GS of Code 2016 is not a bad idea. We can try. Cooool~
On Fri, Mar 4, 2016 at 8:51 PM Neth Six [email protected] wrote:
This is going to happen in Google Summer of Code 2016, or/and probably
earlier (not sure if GSoC requires students to be on board before starting).β
Reply to this email directly or view it on GitHub
#50 (comment)
.
Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
from tensorflow.
it looks like an early stage proposal. we should contact the sciruby team to make this a reality then.
from tensorflow.
Hey @jtoy . It is early stage. While we're awaiting applications from students to develop the Ruby tensorflow API, we're also researching into the best ways to do this. Our first steps would be to enable people to create a tensorflow graph using constants (scalar, matrices), and operations (add, multiply, etc.). Something along the lines of https://www.tensorflow.org/versions/r0.7/get_started/basic_usage.html#basic-usage
from tensorflow.
I was waiting for this ^. I want contribute too(I have a intermediate level of ruby programming skill) but I believe I can do it(help)
from tensorflow.
Dr. Kim,
It's amazing to know that you are employing Tensorflow for real life
purposes!
You should write about it.
Using sklearn is an exercise in Udacity's Deep Learning course. I think it
helps people to get started easier.
Do help SciRuby develop this if you have time.
On Thu, Apr 21, 2016 at 12:57 AM, Sungjin Kim [email protected]
wrote:
Cool! I hope it would be accepted in oder that Ruby users can much enjoy
deep learning with the simple interfaced and powerful Tensorflow package. I
am now using it everyday for researching like medical drug discovery.
Especially I am using Tensorflow with Skflow which is renamed recently as
'learning' underneath of Tensorflow umbrella. So, I hope Ruby-Tensoflow may
include Skflow interface as well. What do you guys think?On Sat, Apr 16, 2016 at 9:50 PM Neth Six [email protected] wrote:
Tensorflow Ruby API is one of the projects submitted to GSoC 2016 by
SciRuby. It is undergoing the selection process, and the results will be
out soon (22/23 Apr).On Sun, Apr 17, 2016 at 2:18 AM, Riley Usagi [email protected]
wrote:What about now?
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<https://github.com/tensorflow/tensorflow/issues/50#issuecomment-210860945>
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<
https://github.com/tensorflow/tensorflow/issues/50#issuecomment-210937423>Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
Definitely. I am happy to help.
On Wed, Apr 20, 2016 at 9:26 PM Neth Six [email protected] wrote:
Dr. Kim,
It's amazing to know that you are employing Tensorflow for real life
purposes!
You should write about it.
Using sklearn is an exercise in Udacity's Deep Learning course. I think it
helps people to get started easier.
Do help SciRuby develop this if you have time.On Thu, Apr 21, 2016 at 12:57 AM, Sungjin Kim [email protected]
wrote:Cool! I hope it would be accepted in oder that Ruby users can much enjoy
deep learning with the simple interfaced and powerful Tensorflow
package. I
am now using it everyday for researching like medical drug discovery.
Especially I am using Tensorflow with Skflow which is renamed recently as
'learning' underneath of Tensorflow umbrella. So, I hope Ruby-Tensoflow
may
include Skflow interface as well. What do you guys think?On Sat, Apr 16, 2016 at 9:50 PM Neth Six [email protected]
wrote:Tensorflow Ruby API is one of the projects submitted to GSoC 2016 by
SciRuby. It is undergoing the selection process, and the results will
be
out soon (22/23 Apr).On Sun, Apr 17, 2016 at 2:18 AM, Riley Usagi <[email protected]
wrote:
What about now?
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<https://github.com/tensorflow/tensorflow/issues/50#issuecomment-210860945>
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<https://github.com/tensorflow/tensorflow/issues/50#issuecomment-210937423>
Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
<
https://github.com/tensorflow/tensorflow/issues/50#issuecomment-212488089>β
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#50 (comment)
Sungjin (James) Kim, PhD
Postdoc, CCB in Harvard
[Web] http://aspuru.chem.harvard.edu/james-sungjin-kim/
[Linkedin] https://www.linkedin.com/in/jamessungjinkim
[Facebook] https://www.facebook.com/jamessungjin.kim
[alternative email] [email protected]
from tensorflow.
@nethsix judging from https://summerofcode.withgoogle.com/organizations/5131000665341952/ it seems that the TensorFlow Ruby API project has not been accepted in GSoC? Am I right?
from tensorflow.
Yes that is correct. We had some applicants, and one of them worked with us
to submit a detailed plan but unfortunately also had some other internship
commitments which raised some concerns.
On Mon, May 2, 2016 at 1:27 PM, jjgh [email protected] wrote:
@nethsix https://github.com/nethsix judging from
https://summerofcode.withgoogle.com/organizations/5131000665341952/ it
seems that the TensorFlow Ruby API project has not been accepted in GsoC?
Am I right?β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
That's so sad =(
from tensorflow.
Hi @nethsix can we take a look at the detailed plan? Is it possible to make it available to us?
from tensorflow.
I mentioned before, but I'd like to help push this forward by sponsoring part of it. We could put together a small fund or request to get this done. I would really like to see this down over the summer. Anyone interested in working on this?
-- Sent from my mobile
On May 2, 2016, at 2:55 AM, Neth Six [email protected] wrote:
Hey Akshay,
GSoC requires students to specify the deliverables each week for the entire
GSoC period, e.g., which API will be wrapped, created, etc. This enables
mentors to track progress accurately, and ensure students stick to their
commitments. I am not sure if it is permissible to share the plan wholesale
but if you think about it, this basically involves looking at Tensorflow
docs, tutorials, etc. to see what is the minimum amount of API to get
started, e.g., model a simple linear regression y = W.x + b requires
tf.placeholder, tf.Variable, tf.matmul, etc. and work from there. The
applicant has also created a repo to wrap the version API, i.e., return the
Tensorflow version, which may seem simple, but it acts a good foundation to
build from. Let me ask the student to share it, and maybe SciRuby can fork
it so that everyone can contribute from there. How does that sound? Any
suggestions?khor
On Mon, May 2, 2016 at 3:15 PM, Akshay Khole [email protected]
wrote:Hi @nethsix https://github.com/nethsix can we take a look at the
detailed plan? Is it possible to make it available to us?β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#50 (comment)β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
from tensorflow.
Hey @jtoy. Let me ask SciRuby about sponsoring the program or would you
prefer to do the sponsorship directly? We can also probably connect with
the student who applied for the GSoC.
On Mon, May 2, 2016 at 8:02 PM, jtoy [email protected] wrote:
I mentioned before, but I'd like to help push this forward by sponsoring
part of it. We could put together a small fund or request to get this done.
I would really like to see this down over the summer. Anyone interested in
working on this?-- Sent from my mobile
On May 2, 2016, at 2:55 AM, Neth Six [email protected] wrote:
Hey Akshay,
GSoC requires students to specify the deliverables each week for the
entire
GSoC period, e.g., which API will be wrapped, created, etc. This enables
mentors to track progress accurately, and ensure students stick to their
commitments. I am not sure if it is permissible to share the plan
wholesale
but if you think about it, this basically involves looking at Tensorflow
docs, tutorials, etc. to see what is the minimum amount of API to get
started, e.g., model a simple linear regression y = W.x + b requires
tf.placeholder, tf.Variable, tf.matmul, etc. and work from there. The
applicant has also created a repo to wrap the version API, i.e., return
the
Tensorflow version, which may seem simple, but it acts a good foundation
to
build from. Let me ask the student to share it, and maybe SciRuby can
fork
it so that everyone can contribute from there. How does that sound? Any
suggestions?khor
On Mon, May 2, 2016 at 3:15 PM, Akshay Khole [email protected]
wrote:Hi @nethsix https://github.com/nethsix can we take a look at the
detailed plan? Is it possible to make it available to us?β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
<
https://github.com/tensorflow/tensorflow/issues/50#issuecomment-216113619>β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHubβ
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
@jtoy This seems like a wonderful opportunity and i would like to work on this. I had applied to Gsoc this year for Supervised machine learning project in sciruby. I was close to being selected but the slots were limited so didn't make it.
Anyways, I am very familiar with the code base of Sciruby projects such as nmatrix, Daru and I have also made contributions to Sciruby and some other python projects too. Please, take a look at my application for Supervised machine learning gem here.
It appears to me that this is a challenging task, but if this could be made a sponsored project and a member of sciruby or Tensor flow could mentor this. Then I would be more than happy to work on this.
from tensorflow.
If you are cool with it. I could write an application for the project, similar to a gsoc application.
from tensorflow.
yeah, an application is great, how much does Gsoc sponsor? I want to figure out how much this would cost and whether or not we need to find other sponsors.
from tensorflow.
@jtoy In Gsoc a total stipend of 5500 dollars is given to the student and 500 dollars to the mentoring organization.
As for this project, I think the sciruby community would be very welcoming to this idea. And if some mentors are free they could even provide some valuable insights for this project.
from tensorflow.
Hey Arafat, @jtoy,
Arafat, I was talking to Sameer (one of the SciRuby admin) about this
earlier and he told me that he has reached out to you.
@jtoy, there is another student who submitted a GSoC proposal for
Tensorflow Ruby API. I've reached out to him as well.
I spoke to Sameer about @jtoy's sponsorship, and basically it may be best
for @jtoy to decide on the deliverables and deadlines, etc. SciRuby would
be more than happy to provide technical assistance required.
On Tuesday, May 3, 2016, Arafat [email protected] wrote:
@jtoy https://github.com/jtoy In Gsoc a total stipend of 5500 dollars
is given to the student and 500 dollars to the mentoring organization.As for this project, I think the sciruby community would be very welcoming
to this idea. And if some mentors are free they could even provide some
valuable insights for this project.β
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
@v0dro I think that's a great idea. Besides the GSoC students, a lot of other people are willing to contribute to this project too. This will help us get a direction on how to proceed and catch up with the progress.
from tensorflow.
Yes we'll be posting an official update soon.
from tensorflow.
It will be great to see this, the Machine Learning challenge site,
crowdai.org, is written in Ruby :-)
On 10 May 2016 at 21:41, Sameer Deshmukh [email protected] wrote:
Hello,
Jason (@jtoy https://github.com/jtoy on GitHub) has volunteered to
sponsor a student to write a Ruby TensorFlow wrapper over 3 months this
summer. SciRuby will be providing technical assistance in this endeavor andPeople interested in winning this grant should go over this blog entry for
further details: http://www.somatic.io/blog/tensorflow-is-coming-to-rubyApplicants are expected to provide a detailed proposal about the approach
they will employ for writing the Ruby wrappers and provide monthly timeline
over the next 3 months starting 1st of June. Deliverables should be on a
monthly basis. We expect a weekly time line and proper evaluation of the
approaches that you think are the best for this project.Proposals will be accepted till 27 May. Please make sure you send us your
proposals before this date.The sponsorship amount will be paid in monthly installments subject to
completion of work for that particular month. The payment will be delayed
if the applicant cannot deliver the promised material.We also looking for additional sponsors for this project so interested
people can get in touch by mailing me at [email protected].Do spread the word!
β
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub
#50 (comment)
from tensorflow.
+1 , and if TensorFlow not support Go, Java, Lua, Javascript, or R. ?
from tensorflow.
I'll also be willing to work on a Ruby-wrapper for Tensorflow!
from tensorflow.
Get in touch with @Arafatk
from tensorflow.
Thanks @Arafatk ! Nice to hear.
From a cursory look, it does seem that this is building ontop of c_api.h using the approach outlined in the how-to, which is great.
If you have concerns around the C API or the how-to, please do let us know with a separate issue (since those files are meant to be useful across all language bindings).
from tensorflow.
To add to @Arafatk's comment above, if there are any Rubyers out there with extra time, feel free to pick up any outstanding issues from our Milestone 0.1.0: https://github.com/somaticio/tensorflow.rb/issues?q=is%3Aopen+is%3Aissue+milestone%3A0.1.0. Some of the issues don't require much knowledge of Tensorflow, but it's a good way to start learning about the inner workings of Tensorflow.
from tensorflow.
fyi - I built this script to spit out ruby files (en masse) from tensor flow proto buffers for use with grpc https://gist.github.com/johndpope/5d176f4eebeb7ec983fa77d945c18fb1
If you want to discuss further -> jump on #grpc channel
https://tensor-flow-talk-invite.herokuapp.com/invite
from tensorflow.
Neat. Thanks for sharing!
On Sat, Nov 5, 2016 at 12:47 PM, John Pope [email protected] wrote:
fyi - I built this script to spit out ruby files (en masse) from tensor
flow proto buffers for use with grpc https://gist.github.com/johndpope/
5d176f4eebeb7ec983fa77d945c18fb1If you want to discuss further -> jump on #grpc channel
https://tensor-flow-talk-invite.herokuapp.com/inviteβ
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#50 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABUdk8bPry9DhCVpFT4DuYnznd7qPi8Mks5q6_w3gaJpZM4GfDTL
.
from tensorflow.
@johndpope I think it should just be https://tensor-flow-talk-invite.herokuapp.com/ sorry for nitpicking :)
from tensorflow.
Hey you guys should join the SciRuby slack channel too.
Here's the link to join: https://sciruby-slack.herokuapp.com/
from tensorflow.
@Arafatk how's the Ruby Asso. Grant work coming along? Can you post your objectives for the grant period here so everyone has a clear idea of what you'll be doing?
from tensorflow.
I will close this bug, since it is not something that would live in the core tensorflow repository. If someone makes a Ruby API we'd be happy to link to it from tensorflow.org.
from tensorflow.
@martinwicke I am the lead developer of Tensorflow.rb which is the Ruby API for Tensorflow.
I have seen a lot of support from the Ruby community for developing this and I have been funded by @jtoy somaticio and also received the Ruby grant.
I have also contacted @asimshankar and @jhseu who have been very helpful with my queries related to the issues that I faced in the development of tensorflow.rb.
If you can link tensorflow.rb from tensorflow.org that would be cool.
Thanks
from tensorflow.
@Arafatk : it's already linked to from https://www.tensorflow.org/api_docs/ :)
from tensorflow.
@asimshankar Thanks you so much. This is great.
from tensorflow.
@asimshankar Thanks you so much. This is great.
what is great?
and why?
from tensorflow.
and what is "is" π‘
from tensorflow.
@LemonAndroid My repo https://github.com/somaticio/tensorflow.rb has been linked from the official docs.
That's great.
from tensorflow.
@Arafatk oh, awesome, legoooooo
from tensorflow.
@Arafatk https://github.com/artyomche9/object_detection/blob/43a98307ab7f19861455acb2ee1441409f5636b7/app/models/image.rb @artyomche9
from tensorflow.
Related Issues (20)
- ImportError: DLL load failed while importing _pywrap_tensorflow_internal: A dynamic link library (DLL) initialization routine failed. HOT 14
- tf.signal.rfft crashes when fft_length is [0] HOT 2
- With the same input and parameter settings, there is a large difference in the output of Dense layer on GPU and CPU.
- The outputs of conv1d differs significantly on the CPU and GPU HOT 3
- tf.raw_ops.Round outputs zeros for any integer tensor HOT 1
- When using exponential as the activation function, the outputs of the CPU and GPU have large differences HOT 2
- With the same input and parameter settings, there is a large difference in the output of LayerNormalization layer on GPU and CPU HOT 7
- Resizing the input shape in TensorFlow Lite C API HOT 5
- The difference in performance on the parameter activation='exponential' is too large HOT 2
- Adam optimiser fails because of negative v_hat values leading to nan in weight update HOT 2
- [TFLite] Could log level control be added to the TFLite C API?
- tf.math.floordiv produces incorrect result when the denominator is `-inf` HOT 1
- Pose missmatch HOT 1
- ValueError: Device /job:localhost/replica:0/task:0/device:CPU:0 is not found HOT 2
- tf.keras.layers.UpSampling2D outputs different output type for input unsigned integer tensors HOT 4
- tf.raw_ops.FakeQuantWithMinMaxVarsPerChannel raises a program abort HOT 1
- resolve code compatibility with my system HOT 3
- tf.linalg.solve single precision wrong result HOT 6
- Tensorflow/Keras_nlp bug HOT 1
- TFLite Hexagon Delegate not working for QCOM SM6375 SOC.
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
π Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. πππ
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google β€οΈ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tensorflow.