Minor patch release
App
Changed
Fabric
Changed
- Let
TorchCollective
works on thetorch.distributed
WORLD process group by default (#16995)
Fixed
- fixed for all
_cuda_clearCublasWorkspaces
on teardown (#16907) - Improved the error message for installing tensorboard or tensorboardx (#17053)
PyTorch
Changed
- Changed to the
NeptuneLogger
(#16761):- It now supports neptune-client 0.16.16 and neptune >=1.0, and we have replaced the
log()
method withappend()
andextend()
. - It now accepts a namespace
Handler
as an alternative toRun
for therun
argument. This means that you can call it likeNeptuneLogger(run=run["some/namespace"])
to log everything to thesome/namespace/
location of the run.
- It now supports neptune-client 0.16.16 and neptune >=1.0, and we have replaced the
- Allow
sys.argv
and args inLightningCLI
(#16808) - Moveed HPU broadcast override to the HPU strategy file (#17011)
Depercated
- Removed registration of
ShardedTensor
state dict hooks inLightningModule.__init__
withtorch>=2.1
(#16892) - Removed the
lightning.pytorch.core.saving.ModelIO
class interface (#16974)
Fixed
- Fixed
num_nodes
not being set forDDPFullyShardedNativeStrategy
(#17160) - Fixed parsing the precision config for inference in
DeepSpeedStrategy
(#16973) - Fixed the availability check for
rich
that prevented Lightning to be imported in Google Colab (#17156) - Fixed for all
_cuda_clearCublasWorkspaces
on teardown (#16907) - The
psutil
package is now required for CPU monitoring (#17010) - Improved the error message for installing tensorboard or tensorboardx (#17053)
Contributors
@awaelchli, @belerico, @carmocca, @colehawkins, @dmitsf, @Erotemic, @ethanwharris, @kshitij12345, @Borda
If we forgot someone due to not matching commit email with GitHub account, let us know :]