Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Sameroom
    @sameroom-bot
    [Paavini, chainer] Is Chainer participating in GSoC? its name isn't visible on GSoC list. Can someone point me to the link
    Sameroom
    @sameroom-bot
    [Shreyas Kowshik, chainer] Hi, I am Shreyas Kowshik, a second year undergraduate majoring in Mathematics and Computer Science at IIT Kharagpur. I would like to contribute to the project of 'Expanding the knowledge base'. Currently I am getting familiar with chainer and getting started with writing basic models using it. I have experience with Deep Learning and have authored a paper on the same.
    Sameroom
    @sameroom-bot
    [Neha Saini, chainer] @asi1024
    [Neha Saini, chainer] hi again, i posted my message in beginner channel,
    [Neha Saini, chainer] I am CS student from india and have experience with python and related libraries used for machine learning or data analysis tasks, like numpy, pandas or spacy or gensim and to be honest i have very little experience with CUDA libraries although have used them underneath with pytorch, but i am very intrested in learning them and was interested in cupy idea as it provides required motivation to dive down into GPU libraries and frameworks can somebody please point out from where I should start if i wish to apply for this project, and i can say I have good experience and programming in python
    Sameroom
    @sameroom-bot
    [Shivansh Beohar, chainer] @crissman Where shall I discuss my doubts regarding the proposals ?
    Sameroom
    @sameroom-bot
    [Yasasvi Yalamarthy, chainer] Hi, My name is Yasasvi and I'm from IIIT Bangalore, India. I am currently pursuing CSE 3rd year and I have a lot of experience in python and its libraries. Past one year I was working in Data Science so I have a pretty knowledge about libraries like numpy, scikit-learn, scipy, pandas, matplotlib, opencv etc. And I'm interested in projects related to cupy and chainer. I had gone through the previous threads posted here. I would like to know who is the mentor for cupy and chainer related projects.
    Sameroom
    @sameroom-bot
    [Shivansh Beohar, chainer] @crissman I want to work on the issue chainer/chainer#6447, I have tried sphinx examples and have gone through the issue as well, Can you please assign it to me ?
    Sameroom
    @sameroom-bot
    [Shivansh Beohar, chainer] I am having trouble building the docs, traceback : https://pastebin.com/pkncUB6D
    [Shivansh Beohar, chainer] I havent installed chainerx, I dont have a GPU, chainer is built successfully and working (win10)
    [Shivansh Beohar, chainer] terminal log here : https://pastebin.com/iNub0x0z
    [Shivansh Beohar, chainer] can someone please help
    [Shivansh Beohar, chainer] I suspect not having a gpu is creating a problem in testing a doc
    Sameroom
    @sameroom-bot
    [nishino, chainer] Building docs in Windows is currently not officially tested, and from the log, it seems the script fails to find the root directory of Chainer source tree.
    The easiest solution I think is to install Chainer in "editable" mode from git source tree, i.e. pip install -e /path/to/chainer/git/root (note -e).
    There's no need of GPUs (nor chainerx) to build docs.
    Sameroom
    @sameroom-bot
    [Shivansh Beohar, chainer] Thanks @nishino That worked like a charm
    [nishino, chainer] +1
    Sameroom
    @sameroom-bot
    [Shivansh Beohar, chainer] @nishino can you please have a look at this https://github.com/chainer/chainer/issues/6447#issuecomment-473289936
    Sameroom
    @sameroom-bot
    [Sazid ali, chainer] Hi my name is Sazid Ali
    I'm a student
    I want to contribute in the FFT Functions Expansion for CuPy project
    Can anyone tell me how can I get started and and where to do the research or keypoints regarding to it
    ....
    Sameroom
    @sameroom-bot

    [SANMITRA DHARMAVARAPU, chainer] Hi all,
    I am SANMITRA from India.
    I want to contribute to Chainer through GSoC
    on the idea Implement LSTM on ChainerX

    Can you please help in getting started:$

    Sameroom
    @sameroom-bot
    [GCS Living Stone, chainer] Hello everyone,
    [GCS Living Stone, chainer] My name is GCS Living Stone. And I have gone through the wiki page of GSoC 2019. I know C++, Python and Machine learning and I want to implement LSTM on ChainerX. Could anyone help me understand the beginners's first steps so that I can follow up with it and make proposals accordingly
    Sameroom
    @sameroom-bot
    [GCS Living Stone, chainer] Looking forward for hearing and taking the guidance.
    Sameroom
    @sameroom-bot
    [Seiya Tokui, chainer] It's good to start from trying some easy ops like elementwise one. There is an issue for adding ops chainer/chainer#6423.
    Sameroom
    @sameroom-bot
    [Kurt, chainer] Hi, I have a question about loading a snapshot generated using the trainer (in contrast to the manual training loop). Here's the code I used and it saved the snapshot. trainer.extend(extensions.snapshot(filename='snapshot_epoch-{.updater.epoch}'), trigger=training.triggers.MaxValueTrigger(key="validation/main/bleu_sen")) However, when I try to load it, I get this error: File "../seq2seq/seq2seq.py", line 856, in main
    serializers.load_npz(args.model, model)
    File "/usr/local/lib/python3.5/dist-packages/chainer/serializers/npz.py", line 190, in load_npz
    d.load(obj)
    File "/usr/local/lib/python3.5/dist-packages/chainer/serializer.py", line 83, in load
    obj.serialize(self)
    File "/usr/local/lib/python3.5/dist-packages/chainer/link.py", line 997, in serialize
    d[name].serialize(serializer[name])
    File "/usr/local/lib/python3.5/dist-packages/chainer/link.py", line 651, in serialize
    data = serializer(name, param.data)
    File "/usr/local/lib/python3.5/dist-packages/chainer/serializers/npz.py", line 143, in __call_
    dataset = self.npz[key]
    File "/usr/local/lib/python3.5/dist-packages/numpy/lib/npyio.py", line 255, in
    _getitem__
    raise KeyError("%s is not a file in the archive" % key)
    KeyError: 'entangle_h2/b is not a file in the archive'
    Sameroom
    @sameroom-bot
    [Kurt, chainer] But when I printed the model.keys, I can find that key there: updater/model:main/entangle_h2/b This is my first of using the trainer to save a snapshot and when I was using the manual training loop, this did not happen. Does anybody know how to resolve this, please? Cheers
    Sameroom
    @sameroom-bot
    [Crissman Loomis, chainer] Deadline for GSoC submissions is April 10! Please let us know if you’re considering submitting for Chainer or CuPy.
    Sameroom
    @sameroom-bot
    [Shresth Verma, chainer] Hi Crissman,
    [Shresth Verma, chainer] Should the proposal be added in the Chianer repo's wiki? Or is it just fine to put it on drive and share the draft through official GSoC website?
    Sameroom
    @sameroom-bot
    [Rishav Chourasia, chainer] Hi. I am Rishav Chourasia, a PhD candidate at NUS school of computing starting this fall. I want to contribute towards expanding chainer compiler coverage as part of GSoC. I have been contributing to Optuna hyperparameter framework for a couple of weeks. I have some questions regarding the project.
    [Rishav Chourasia, chainer] 1) I need some starting pointers for understanding the experimental chainer compiler, like where to start from and what major libraries I need to be familiar with.
    2) An example of a recent expansion added to the compiler so that I can follow what is the expectation for the project.
    Sameroom
    @sameroom-bot
    [Shresth Verma, chainer] Hi!
    I have shared my proposal draft for the Expanding Chainer Compiler Coverage project through the GSoC website. It would be really great if one of the mentors (suggested mentor: Shinichiro Hamaji) could suggest some reviews.
    Sameroom
    @sameroom-bot
    [Akshay Subramanian, chainer] My draft proposal for Expand ChainerX Ops has also been submitted through the GSOC portal. I would love getting some feedback too. (Project mentors: @beam2d, @nishino, @hvy, @imanishi)
    Sameroom
    @sameroom-bot
    [Piyush Raikwar, chainer] I've shared my draft proposal for Cupy automatic fallback to Numpy on GSoC portal. @imanishi Could you review it? And it'd be great to get some suggestions.
    Sameroom
    @sameroom-bot
    [Vikram Chaurasia, chainer] Hi, I've shared my draft proposal for CuPy coverage of NumPy functions on GSoC website. It would be my pleasure if one of the mentors (suggested: @imanishi) would review it and give some suggestions.
    Sameroom
    @sameroom-bot
    [Harshan Baskar, chainer] Hi! I have shared my draft proposal for /Expanding ChainerX ops/ on GSoC portal. I would like get some feedback regarding it. (Mentors: @beam2d, @nishino, @hvy, @imanishi)
    Sameroom
    @sameroom-bot
    [Rishav Chourasia, chainer] I too have shared my draft proposal for Expanding Chainer Compiler Coverage on GSoC portal. Please provide feedback for on it as well. @hamaji I would like @hamaji to be my mentor in the project.
    Sameroom
    @sameroom-bot

    [Apoorv Jain, chainer] Hey Everyone!

    I am Apoorv Jain, a final-year undergraduate student of Computer Science & Engineering from Bharati Vidyapeeth's College of Engineering. I have knowledge & experience of working on various python modules such as OpenCV, Scikit-learn, Scipy, NumPy, MatPlotlib, Pandas, Chatterbot, Flask, NLTK and, scrapy etc. Along with it in the domain of deep learning, I have experience of working on Keras, Tenserflow and great experience with transfer learning as well. I even worked on several Image processing projects out of my curriculum.

    I am interested in contributing to Chainer as it has some really great projects to work upon and also they align with my experience. I am mainly interested in "Expand the Knowledge Base".

    These projects are based on Python, which is my domain of interest. Kindly check out my profile on LinkedIn for more information: https://www.linkedin.com/in/apoorv-jain-110769100/

    Github: https://github.com/apoorvjain25

    Kindly let me know how to proceed forward with the application.

    Sameroom
    @sameroom-bot
    [Akshay Subramanian, chainer] I have submitted my draft proposal for the project Expand Chainer Primitive Operations through the GSOC portal as well. I'd like to get some feedback and suggestions on this proposal too. (Project mentors: @beam2d, @nishino, @hvy, @imanishi)
    Sameroom
    @sameroom-bot
    [Ivan Yashchuk, chainer] Hello! I've shared my draft proposal for Expanding ChainerX Ops: Differentiable Linear Algebra on the GSoC website. It would be great to receive feedback on that.
    Sameroom
    @sameroom-bot
    [Aniket Didolkar, chainer] Hi! I have shared my draft proposal titled Recurrent Neural Networks for ChainerX through the GSOC portal. I would love some feedback on the proposal. (Project mentors: @hvy, @beam2d)
    Sameroom
    @sameroom-bot
    [Harshan Baskar, chainer] Hi. I have shared the draft of my proposal for ChainerX C++ API Tutorial. Any feedback on the draft would be greatly helpful. (Mentors: @beam2d, @imanishi, @hvy, @nishino)
    Sameroom
    @sameroom-bot
    [Crissman Loomis, chainer] Thanks to all of you who have made submissions for GSoC! We’ve looked over them all, although we haven’t necessarily given comments on all of them.
    Sameroom
    @sameroom-bot
    [Crissman Loomis, chainer] here There is less than 16 HOURS left until the submission deadline. The status of the proposal must be FINAL in order for us to be able to select it. Please make sure to mark your proposal as FINAL, even if we haven’t given specific feedback.
    Sameroom
    @sameroom-bot
    [Crissman Loomis, chainer] Thanks to those who made submissions. We’re currently reviewing, and next step is to get back to NUMFocus with our minimum and maximum number of slots desired.
    Sameroom
    @sameroom-bot
    [Satoshi Tsutsui, chainer] Does chainer support gradient accumulation? I'd like to train a model with large batch size using only one GPU. The model is large but still fits into a GPU memory if batch size is 1. So i want to run forward computation several times, then gradually accumulate gradients until it becomes equivalent to a large batch size, then finally update the model parameters.
    [Seiya Tokui, chainer] If you run multiple forward-backward without resetting the parameter gradients, Chainer automatically accumulates the gradients.
    [Seiya Tokui, chainer] So you can emulate large batch training by multiple backprops with small batchsize.
    Sameroom
    @sameroom-bot
    [Satoshi Tsutsui, chainer] Wow! Thanks for the reply! That's great to know!
    Sameroom
    @sameroom-bot
    [dhgrs, chainer] If the network includes batch normalization or other normalization, it does not work well, right?
    Sameroom
    @sameroom-bot
    [Ishan Rai, chainer] I am getting the following error while running test after recently pulled from the master. The tests are successful on older branches: