Re: Is ML module @IgniteExperimental?
No, that is right! вс, 12 апр. 2020 г., 5:14 Maxim Muzafarov : > Folks, > > It seems to me we've forgotten to remove -Ptensorflow profile from the > ~Build Apache Ignite~ suite [1] on the project build. > I've removed it. Please, revert it back if I'm missing something > during this discussion. > > [1] > https://ci.ignite.apache.org/admin/editBuildRunners.html?id=buildType:IgniteTests24Java8_BuildApacheIgnite > > On Wed, 25 Mar 2020 at 02:21, Denis Magda wrote: > > > > Alexey, thanks for sharing details and your reasoning behind the taken > > actions. It makes sense. I've updated the machine learning pages on the > new > > website that will be released in several days. > > > > - > > Denis > > > > > > On Tue, Mar 24, 2020 at 11:07 AM Alexey Zinoviev > > > wrote: > > > > > Hi, Denis! > > > > > > Be honest, the significant amount of the ML contirbutors left the > community > > > previous year in frustration with unfinished parts. > > > In this situation, I reduced the unsed and broken parts according our > > > previous discussions peer-to-peer (not on devlist, our mistake) to > release > > > the stable core of ML which could be supported with reduced power. > > > > > > The reasons for GA removal > > > 1. It doesn't related to the ML topic > > > 2. It has no intersection with the ML package (as you mentioned) > > > 3. It doesn't support Ignite code and in many places Java codestyle > > > 4. It was experimental package placed in ML in time of earliest > experiments > > > in 2017 > > > 5. Nobody doesn't want to support this for the years > > > > > > Genetic Algorithms could be moved to Ignite-extension (if somebody > > > interested in it) > > > > > > A lot of things are changed since release 2.7 > > > > > > Lessons are learnt, I will start discussion topics next time for the > > > significant changes or removal in API, moreover, the next releases I > hope > > > to use new @IgniteExperimental (it was added too late) and another > > > annotations for the release cycle. > > > > > > вт, 24 мар. 2020 г. в 20:00, Denis Magda : > > > > > > > Alexey, > > > > > > > > I missed this thread and only now realized that TensorFlow, genetic > > > > algorithms and some other APIs were expelled from 2.8. I would > encourage > > > us > > > > to start a dedicated discussion for any APIs removal or significant > > > changes > > > > to let other community members share their opinions or take > appropriate > > > > actions (like proper documentation redirects setup for pages that are > > > gone > > > > and updates on the website like [1] and [2]). For instance, I have no > > > glue > > > > that the topic of TensorFlow removal was briefly mentioned in this > > > > discussion thread. > > > > > > > > I see the reasoning about TensorFlow but why have we removed generic > > > > algorithms that had a dependency on the compute APIs only? > > > > > > > > [1] https://ignite.apache.org/features/tensorflow.html > > > > [2] https://ignite.apache.org/features/machinelearning.html#ga-grid > > > > > > > > - > > > > Denis > > > > > > > > > > > > On Mon, Feb 17, 2020 at 5:39 AM Alexey Zinoviev < > zaleslaw@gmail.com> > > > > wrote: > > > > > > > > > Ok, agree, that I should start discussion before making changes, > but I > > > > was > > > > > limited by release 2.8 and trying don;'t be a delayed person for > that. > > > > > During release I was focused on fixing bugs and don't tests TF and > > > Ignite > > > > > together > > > > > > > > > > I thought that as a maintainer of ML module I could do perform > these > > > > > actions. > > > > > > > > > > Below I will share my statement why it should be removed and why it > > > > should > > > > > be removed immediately > > > > > > > > > > About TensorFlow module (reason for removal) > > > > > > > > > >1. This module is only one module that uses IGFS and needs in > > > > FileSystem > > > > >on Ignte side due to TensorFlow API > > > > >2. This module a part of bridge between Ignite ML and > Tensorflow and > > > > its > > > > >broken after changes in TensorFlow on TensorFlow side > > > > >3. TensorFlow released new version without Ignite bridge, no > chance > > > to > > > > >run them together for new releases > > > > >4. This module wasn't complete and developer who did this, left > the > > > > >community > > > > >5. The development skills for this story require python/C++/java > > > > >programming together > > > > >6. The module is a source of bugs which could be fixed for > release > > > 2.8 > > > > >and possibly for future releases (nobody in community could > this) > > > > >7. The release size reduced from 6 Gb to 4.5 due to removed > > > > dependencies > > > > >8. TensorFlow now is not popular among Data Scientists, the > PyTorch > > > is > > > > >the most popular tool for Deep Learning (like NetBeans and IDEA) > > > > >9. Nobody uses that in production because it was developed > between > > > 2.7 > > > > >and 2.8 (2.7 has only
Re: Is ML module @IgniteExperimental?
Folks, It seems to me we've forgotten to remove -Ptensorflow profile from the ~Build Apache Ignite~ suite [1] on the project build. I've removed it. Please, revert it back if I'm missing something during this discussion. [1] https://ci.ignite.apache.org/admin/editBuildRunners.html?id=buildType:IgniteTests24Java8_BuildApacheIgnite On Wed, 25 Mar 2020 at 02:21, Denis Magda wrote: > > Alexey, thanks for sharing details and your reasoning behind the taken > actions. It makes sense. I've updated the machine learning pages on the new > website that will be released in several days. > > - > Denis > > > On Tue, Mar 24, 2020 at 11:07 AM Alexey Zinoviev > wrote: > > > Hi, Denis! > > > > Be honest, the significant amount of the ML contirbutors left the community > > previous year in frustration with unfinished parts. > > In this situation, I reduced the unsed and broken parts according our > > previous discussions peer-to-peer (not on devlist, our mistake) to release > > the stable core of ML which could be supported with reduced power. > > > > The reasons for GA removal > > 1. It doesn't related to the ML topic > > 2. It has no intersection with the ML package (as you mentioned) > > 3. It doesn't support Ignite code and in many places Java codestyle > > 4. It was experimental package placed in ML in time of earliest experiments > > in 2017 > > 5. Nobody doesn't want to support this for the years > > > > Genetic Algorithms could be moved to Ignite-extension (if somebody > > interested in it) > > > > A lot of things are changed since release 2.7 > > > > Lessons are learnt, I will start discussion topics next time for the > > significant changes or removal in API, moreover, the next releases I hope > > to use new @IgniteExperimental (it was added too late) and another > > annotations for the release cycle. > > > > вт, 24 мар. 2020 г. в 20:00, Denis Magda : > > > > > Alexey, > > > > > > I missed this thread and only now realized that TensorFlow, genetic > > > algorithms and some other APIs were expelled from 2.8. I would encourage > > us > > > to start a dedicated discussion for any APIs removal or significant > > changes > > > to let other community members share their opinions or take appropriate > > > actions (like proper documentation redirects setup for pages that are > > gone > > > and updates on the website like [1] and [2]). For instance, I have no > > glue > > > that the topic of TensorFlow removal was briefly mentioned in this > > > discussion thread. > > > > > > I see the reasoning about TensorFlow but why have we removed generic > > > algorithms that had a dependency on the compute APIs only? > > > > > > [1] https://ignite.apache.org/features/tensorflow.html > > > [2] https://ignite.apache.org/features/machinelearning.html#ga-grid > > > > > > - > > > Denis > > > > > > > > > On Mon, Feb 17, 2020 at 5:39 AM Alexey Zinoviev > > > wrote: > > > > > > > Ok, agree, that I should start discussion before making changes, but I > > > was > > > > limited by release 2.8 and trying don;'t be a delayed person for that. > > > > During release I was focused on fixing bugs and don't tests TF and > > Ignite > > > > together > > > > > > > > I thought that as a maintainer of ML module I could do perform these > > > > actions. > > > > > > > > Below I will share my statement why it should be removed and why it > > > should > > > > be removed immediately > > > > > > > > About TensorFlow module (reason for removal) > > > > > > > >1. This module is only one module that uses IGFS and needs in > > > FileSystem > > > >on Ignte side due to TensorFlow API > > > >2. This module a part of bridge between Ignite ML and Tensorflow and > > > its > > > >broken after changes in TensorFlow on TensorFlow side > > > >3. TensorFlow released new version without Ignite bridge, no chance > > to > > > >run them together for new releases > > > >4. This module wasn't complete and developer who did this, left the > > > >community > > > >5. The development skills for this story require python/C++/java > > > >programming together > > > >6. The module is a source of bugs which could be fixed for release > > 2.8 > > > >and possibly for future releases (nobody in community could this) > > > >7. The release size reduced from 6 Gb to 4.5 due to removed > > > dependencies > > > >8. TensorFlow now is not popular among Data Scientists, the PyTorch > > is > > > >the most popular tool for Deep Learning (like NetBeans and IDEA) > > > >9. Nobody uses that in production because it was developed between > > 2.7 > > > >and 2.8 (2.7 has only proof-of-concept) > > > > > > > > Nikolay, sorry for that, hope to share more information about the ML > > and > > > > discuss here the main changes before actions. > > > > > > > > пн, 17 февр. 2020 г. в 16:18, Nikolay Izhikov : > > > > > > > > > Hello, Alexey. > > > > > > > > > > > The main reason, the modules are not work proper way, were > > > > experimental,
[jira] [Created] (IGNITE-12888) Add support for ConstantName to checkstyle rules
Maxim Muzafarov created IGNITE-12888: Summary: Add support for ConstantName to checkstyle rules Key: IGNITE-12888 URL: https://issues.apache.org/jira/browse/IGNITE-12888 Project: Ignite Issue Type: Task Reporter: Maxim Muzafarov Assignee: Maxim Muzafarov Add a new rule to checkstyle according to Apache Ignite naming conventions. https://cwiki.apache.org/confluence/display/IGNITE/Coding+Guidelines#CodingGuidelines-Naming -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: [DISCUSSION] Major changes in Ignite in 2020
Folks, I keep working on tasks related to data consistency. This includes: Lost partitions handling overhaul (almost done) and tombstones support aforementioned by Ivan Rakov Atomic protocol overhaul (see [1]) The ultimate goal of the year is to prepare Ignite for passing Jepsen tests. [1] https://cwiki.apache.org/confluence/display/IGNITE/IEP-12+Make+ATOMIC+Caches+Consistent+Again пт, 10 апр. 2020 г. в 18:49, Denis Magda : > Steven, > > Please start a dedicated discussion for the Golang support. At the moment, > I'm not aware if anybody from the community planned to provide support > out-of-the-box. However, that's not a tricky task thanks to Ignite's binary > protocol that lets enable any programming language support easily. > > - > Denis > > > On Fri, Apr 10, 2020 at 8:43 AM smeadows-abb > wrote: > > > First thanks for your quick response. > > > > I looked at https://github.com/amsokol/ignite-go-client and its NOT > > completed and nothing has been for last 16 months. Initial test with > > package > > failed, so trying to determine your project roadmap with regards to > Golang > > and maybe Rust support. > > > > I'm NOT sure of 'AFAIK' ? > > > > We may need to implement your Restful API to provide support for Golang > and > > Rust, provided it's complete? > > > > Thanks, > > Steve > > > > > > > > > > > > -- > > Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/ > > > -- Best regards, Alexei Scherbakov
