pytorch/PiPPy issues and pull requests

#742 - Make output chunk spec optional

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago - 2 comments
Labels: cla signed

#741 - Fix minGPT CI

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#740 - Make input chunk spec optional

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago - 1 comment
Labels: cla signed

#739 - [SPMD] Fix a linter error in distribute

Pull Request - State: closed - Opened by fegin almost 2 years ago
Labels: cla signed

#738 - Fix issue that PiPPy does not generate backward pass when there are multiple output values

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#737 - [spmd] spmd logging doesn't work with logging level

Issue - State: closed - Opened by wanchaol almost 2 years ago

#736 - [SPMD] Add support for FakeTensor mode and avoid running the op dispatch twice

Pull Request - State: closed - Opened by anj-s almost 2 years ago - 1 comment
Labels: cla signed

#735 - [spmd] lint fixes

Pull Request - State: closed - Opened by wanchaol almost 2 years ago
Labels: cla signed

#734 - Revert "[spmd] refactor expansion logic"

Pull Request - State: closed - Opened by anj-s almost 2 years ago
Labels: cla signed

#733 - [SPMD] Dynamo FE for the SPMD API

Pull Request - State: closed - Opened by anj-s almost 2 years ago - 1 comment
Labels: cla signed

#732 - [SPMD] Clean 3/N

Pull Request - State: closed - Opened by anj-s almost 2 years ago
Labels: cla signed

#731 - Pippy ddp2pipe example doesn't work for pipeline

Issue - State: open - Opened by yurishin929 almost 2 years ago - 4 comments

#730 - [SPMD] Cleanup 2/N

Pull Request - State: closed - Opened by anj-s almost 2 years ago
Labels: cla signed

#729 - [checkpoint] Fix pg planner

Pull Request - State: closed - Opened by wz337 almost 2 years ago
Labels: cla signed

#728 - Fix lint errors on main branch

Pull Request - State: closed - Opened by anj-s almost 2 years ago
Labels: cla signed

#727 - [spmd] refactor expansion logic

Pull Request - State: closed - Opened by wanchaol almost 2 years ago - 1 comment
Labels: cla signed

#726 - [SPMD] Cleanup - 1/N

Pull Request - State: closed - Opened by anj-s almost 2 years ago
Labels: cla signed

#725 - [checkpoint] Add example to show how to extend FileSystemWriter

Pull Request - State: closed - Opened by wz337 almost 2 years ago
Labels: cla signed

#724 - [spmd] delete dtensor op db tests

Pull Request - State: closed - Opened by wanchaol almost 2 years ago
Labels: cla signed

#723 - How to reduce memory costs when running on CPU

Issue - State: open - Opened by jiqing-feng almost 2 years ago

#722 - [spmd] delete dtensor op db tests and fixes CI

Pull Request - State: closed - Opened by wanchaol almost 2 years ago
Labels: cla signed

#721 - [DTensor] missing rule for aten.fill.Scalar causing unit tests to fail for SPMD

Issue - State: open - Opened by lessw2020 almost 2 years ago

#720 - [WIP] Capture whole fused graph, perform SPMD expansion, run with Inductor

Pull Request - State: closed - Opened by aazzolini almost 2 years ago - 5 comments
Labels: cla signed

#719 - Example of capturing training step and lowering to inductor

Pull Request - State: closed - Opened by aazzolini almost 2 years ago - 1 comment
Labels: cla signed

#718 - Example of capturing training step and lowering to inductor

Pull Request - State: closed - Opened by aazzolini almost 2 years ago - 1 comment
Labels: cla signed

#717 - JIT fusion, with Memory size fusion policy

Pull Request - State: closed - Opened by lessw2020 almost 2 years ago - 1 comment
Labels: cla signed

#716 - [SPMD] Tighten `_is_param` data ptr check

Pull Request - State: closed - Opened by awgu almost 2 years ago - 6 comments
Labels: cla signed

#715 - Issue with FX tracing of HF seq2seq models

Issue - State: open - Opened by HamidShojanazeri almost 2 years ago

#714 - fixing the example if conditions

Pull Request - State: closed - Opened by HamidShojanazeri almost 2 years ago
Labels: cla signed

#713 - Fix PiPPy + Dynamo CI

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#712 - Fix hf_model_tests

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#711 - Remove torch_required per Transformer upgrade

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#710 - fix barrier params

Pull Request - State: closed - Opened by jiqing-feng almost 2 years ago - 2 comments
Labels: cla signed

#706 - [SPMD] Add support for logging configs

Pull Request - State: closed - Opened by anj-s almost 2 years ago - 1 comment
Labels: cla signed

#698 - Distx deferred init

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#687 - Unify argument parsing across tests and examples

Pull Request - State: closed - Opened by acforvs almost 2 years ago - 27 comments
Labels: cla signed

#682 - GPT2 use all GPUs

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#678 - [SPMD] Populate primals <-> param <-> grad mapping

Pull Request - State: closed - Opened by fegin almost 2 years ago
Labels: cla signed, SPMD

#675 - Adding a issue and PR template

Pull Request - State: closed - Opened by HamidShojanazeri almost 2 years ago
Labels: cla signed

#665 - Disable DDP bucket rebuild to avoid hang in ddp2pipe

Pull Request - State: closed - Opened by kwen2501 almost 2 years ago
Labels: cla signed

#653 - try using aot_module_simplified

Pull Request - State: closed - Opened by wanchaol almost 2 years ago - 2 comments
Labels: cla signed

#631 - Feature remove clones source

Pull Request - State: closed - Opened by lessw2020 about 2 years ago
Labels: cla signed

#626 - add _remove_gradient_tensor_clones

Pull Request - State: closed - Opened by lessw2020 about 2 years ago - 1 comment
Labels: cla signed

#576 - Move DTensor from tau to PyTorch

Issue - State: closed - Opened by wanchaol about 2 years ago
Labels: SPMD

#569 - PiPPy + TP

Pull Request - State: closed - Opened by kwen2501 about 2 years ago
Labels: cla signed

#562 - [spmd] spmd api tracing warning need to investigate

Issue - State: closed - Opened by wanchaol about 2 years ago - 2 comments
Labels: SPMD

#517 - Fix print_readable

Pull Request - State: closed - Opened by yunxing about 2 years ago - 2 comments
Labels: fx

#511 - [spmd] consolidate behavior when rank is not participate in the mesh

Issue - State: closed - Opened by wanchaol about 2 years ago
Labels: SPMD, distributed_tensor

#483 - HF Accelerate examples

Pull Request - State: closed - Opened by pbelevich about 2 years ago - 1 comment
Labels: cla signed

#434 - use torchDynamo instead of tracing to get graph?

Issue - State: open - Opened by Jack47 about 2 years ago - 3 comments
Labels: enhancement, PiPPy

#429 - Split PiPPyTrainingArguments

Pull Request - State: closed - Opened by kwen2501 about 2 years ago
Labels: cla signed

#428 - Factor out `intra_op_costs_estimator` from `AutoParallelConfig`

Pull Request - State: closed - Opened by makslevental about 2 years ago
Labels: cla signed

#411 - CI: pin Huggingface version to bring CI back to green

Pull Request - State: closed - Opened by kwen2501 over 2 years ago
Labels: cla signed

#410 - CI: switch to T4 GPU from Maxwell

Pull Request - State: closed - Opened by kwen2501 over 2 years ago
Labels: cla signed

#408 - [spmd] sharding propagation should handle reshard suggestion with care

Issue - State: closed - Opened by wanchaol over 2 years ago
Labels: SPMD

#347 - [spmd] add N-D tests for point-wise ops

Issue - State: closed - Opened by wanchaol over 2 years ago
Labels: SPMD

#335 - update mm rules to be more generic

Issue - State: closed - Opened by wanchaol over 2 years ago
Labels: SPMD

#326 - [spmd] device mesh to allow user to run multi-dim together

Issue - State: closed - Opened by wanchaol over 2 years ago
Labels: SPMD

#315 - Test RPC meta device

Pull Request - State: closed - Opened by pbelevich over 2 years ago - 1 comment
Labels: cla signed

#285 - Use torch-mlir and MLIR to estimate the latencies of fx graph ops

Pull Request - State: closed - Opened by makslevental over 2 years ago
Labels: cla signed

#276 - Double pipe scheme

Pull Request - State: closed - Opened by kwen2501 over 2 years ago
Labels: cla signed, PiPPy

#247 - Support LayoutLM models in HF tests

Issue - State: open - Opened by pbelevich over 2 years ago - 1 comment
Labels: good first issue, huggingface, PiPPy

#131 - Update HF GPT-2 test to forward-backward

Pull Request - State: closed - Opened by pbelevich over 2 years ago - 1 comment
Labels: cla signed, PiPPy

#25 - Make RemoteInterpreter use the full implementation of `Interpreter.run`

Issue - State: open - Opened by jamesr66a almost 3 years ago - 2 comments
Labels: good first issue, mid-pri

GitHub / pytorch/PiPPy issues and pull requests