8000 DTensor key word args to PrepareModuleInput not handled correctly · Issue #128365 · pytorch/pytorch · GitHub
[go: up one dir, main page]
Skip to content 8000
DTensor key word args to PrepareModuleInput not handled correctly #128365
New issue
New issue
Closed
Closed
DTensor key word args to PrepareModuleInput not handled correctly#128365
Assignees
wanchaol
Labels
module: dtensordistributed tensor tagoncall: distributedAdd this issue/PR to distributed oncall triage queuetriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
@schinmayee

Description

@schinmayee
schinmayee
opened on Jun 10, 2024

Key word args to PrepareModuleInput can be DTensors, which are not handled correctly here.

pytorch/torch/distributed/tensor/parallel/style.py

Line 438 in 739aa22

kwarg_val = DTensor.from_local(kwarg_val, device_mesh, (input_layout,), run_check=False)

cc @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @osalpekar @jiayisuse @H-Huang @kwen2501 @awgu @penguinwu @fegin @XilunWu @wanchaol @fduwjj @wz337 @tianyu-l @wconstab @yf225 @chauhang @d4l3k

Metadata

Metadata

Assignees

Labels

module: dtensordistributed tensor tagoncall: distributedAdd this issue/PR to distributed oncall triage queuetriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions
    0