Search results

Slower inference performance when switching from torchxla_trace_once to openxla compile backend

dynamo

Status: Open.

#5430 In pytorch/xla;

· GleasonK opened on Aug 9, 2023

Run backward on CPU before importing torch_xla cause future backward on XLA crash

dynamo

nostaleDo not consider for stalenessDo not consider for staleness

Status: Open.

#4174 In pytorch/xla;

· shunting314 opened on Nov 9, 2022

ExecuteTime metric not accurate

dynamo

triagedThis issue has been reviewed by the triage team and the appropriate priority assigned.This issue has been reviewed by the triage team and the appropriate priority assigned.

Status: Open.

#4306 In pytorch/xla;

· shunting314 opened on Dec 9, 2022

[Dynamo] Make _run_cached_graph not blocking for dynamo

dynamo

Status: Open.

#4402 In pytorch/xla;

· JackCaoG opened on Dec 28, 2022

[Dynamo] Support force_ltc_data in GetGraphHash

dynamo

Status: Open.

#4401 In pytorch/xla;

· JackCaoG opened on Dec 28, 2022

Dynamo incorrectly slice the tensor value at 0th dimension

dynamo

Status: Open.

#5548 In pytorch/xla;

· JackCaoG opened on Sep 7, 2023

TPU dynamo speed up on Training analysis

dynamo

Status: Open.

#4475 In pytorch/xla;

· JackCaoG opened on Jan 19, 2023

Dynamo compiled_autograd crashed when leaf tensor is on CPU device and move to xla later

dynamo

Status: Open.

#5676 In pytorch/xla;

· JackCaoG opened on Oct 4, 2023

[torch_xla] scan only captures aten operations

dynamo

enhancementNew feature or requestNew feature or request

SPMD / Distributed

Status: Open.

#8691 In pytorch/xla;

· tengyifei opened on Feb 9, 2025

Two different graph have the same hash value

bugSomething isn't workingSomething isn't working

dynamo

Status: Open.

#8353 In pytorch/xla;

· yitongh opened on Nov 4, 2024