- 17 Sep, 2020 1 commit
-
-
Tom Birch authored
Adds support for distributing pipeline stages across multiple processes (and therefore multiple machines) * Adds a style argument to the Pipe constructor, defaulting to PipelineStyle.SingleProcess, but also supporting PipelineStyle.MultiProcess * Added support for lazy construction of modules (see lazy_construction for an example) * Added two implementations of inter-process communication: one based on rpc with globally visible queues, one based on send/recv * Copied all the relevant tests from tests/pipe to tests/pipe_process and modified them to exercise PipelineStyle.MultiProcess
-
- 16 Sep, 2020 1 commit
-
-
msbaines authored
-
- 03 Sep, 2020 1 commit
-
-
Jun Ru Anderson authored
Add GradScaler to Fairscale, subclassing PyTorch's GradScaler. Use GradScaler in the pipe benchmark; though it is not needed in this case, it is a good example of how to use gradient scaling for larger models that do require gradient scaling in order to converge. Co-authored-by:Jun Ru Anderson <andersonic@fb.com>
-
- 27 Aug, 2020 1 commit
-
-
msbaines authored
Workaround PyTorch bug that casts state (pytorch/pytorch#43706). Copied from https://github.com/pytorch/fairseq/blob/v0.9.0/fairseq/optim/fp16_optimizer.py#L251-L268
-
- 14 Aug, 2020 2 commits
-
-
msbaines authored
authored-by:Mandeep Singh Baines <msb@fb.com>
-
msbaines authored
-
- 31 Jul, 2020 3 commits
-
-
msbaines authored
-
Tom Birch authored
-
Jun Ru Anderson authored
Co-authored-by:Jun Ru Anderson <andersonic@fb.com>
-
- 08 Jul, 2020 1 commit
-
-
Mandeep Singh Baines authored
-