Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / chainer/chainer-pytorch-migration issues and pull requests

#28 - Deprecate cupy memory pool sharing feature

Issue - State: open - Opened by kmaehashi over 4 years ago

#27 - Add `chainer_torch_function` and `TorchCainerFunction`

Pull Request - State: open - Opened by emcastillo over 4 years ago - 1 comment

#26 - Rely on `__cuda_array_interface__` for tensor conversions

Pull Request - State: open - Opened by emcastillo over 4 years ago

#25 - Add example of `TorchModule`

Pull Request - State: open - Opened by emcastillo over 4 years ago

#24 - How to override StandardUpdater.update_core method when using TorchModule class

Issue - State: closed - Opened by take0212 almost 5 years ago - 21 comments

#22 - Fix `cpm.astensor(cupy.ndarray)` for zero-size input

Pull Request - State: closed - Opened by asi1024 almost 5 years ago - 3 comments

#21 - Support `LinkAsTorchModule.to(device)`

Pull Request - State: closed - Opened by asi1024 almost 5 years ago - 4 comments

#20 - Bump version to v0.0.2

Pull Request - State: closed - Opened by kmaehashi almost 5 years ago - 2 comments

#19 - Add `to_chainer_device` and `to_torch_device`

Pull Request - State: closed - Opened by asi1024 almost 5 years ago - 8 comments

#18 - Fix LinkAsTorchModel test for uninitialized parameter

Pull Request - State: closed - Opened by niboshi almost 5 years ago - 4 comments

#17 - flexci: use cupy v7 to be compatible with chainer

Pull Request - State: closed - Opened by toslunar almost 5 years ago - 4 comments

#17 - flexci: use cupy v7 to be compatible with chainer

Pull Request - State: closed - Opened by toslunar almost 5 years ago - 4 comments

#16 - Fix backprop of `LinkAsTorchModel`

Pull Request - State: closed - Opened by niboshi about 5 years ago - 5 comments

#15 - Better error message for uninitialized parameters in LinkAsTorchModel

Pull Request - State: closed - Opened by niboshi about 5 years ago - 5 comments

#14 - Support ChainerMN `multi_node_snapshot`

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 4 comments

#13 - Fix epoch calculation when restoring snapshot in ignite

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 12 comments

#12 - Resumption from snapshot taken in the middle of the epoch

Issue - State: closed - Opened by msakai about 5 years ago - 1 comment

#11 - multi_node_snapshot does not work

Issue - State: closed - Opened by msakai about 5 years ago - 1 comment

#9 - `asarray` support for multiple gpus

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 1 comment
Labels: bug

#8 - `DataParallel` module name conflicts with `TorchModule.module` attr

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 7 comments

#7 - Add custom collate functions for the `DataLoader`

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 4 comments

#6 - Add `forward` method to `LinkAsTorchModel`

Pull Request - State: closed - Opened by emcastillo about 5 years ago - 18 comments

#5 - cpm.tensor.asarray raises CUDARuntimeError when a tensor resides in non-default GPU

Issue - State: closed - Opened by msakai about 5 years ago
Labels: bug

#4 - cpm.ignite.add_trainer_extension not compatible with DataParallel

Issue - State: closed - Opened by msakai about 5 years ago
Labels: bug

#3 - Snapshot extension's autoloading raises AttributeError

Issue - State: open - Opened by msakai about 5 years ago - 2 comments

#2 - Fix typo: snapeshot -> snapshot

Pull Request - State: closed - Opened by msakai about 5 years ago - 1 comment

#1 - Training LinkAsTorchModel-wrapped Chainer model using PyTorch + Ignite?

Issue - State: closed - Opened by msakai about 5 years ago - 7 comments