Search results

Llama3.2 vision model AOTI integration

Compile / AOTIIssues related to AOT Inductor and torch compileIssues related to AOT Inductor and torch compile

enhancementNew feature or requestNew feature or request

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1497 In pytorch/torchchat;

· larryliu0820 opened on Feb 21, 2025

We still don't use BFDOT on macOS CPU because Apple compiler used for PyTorch wheel is outdated

Known GapsThese are known Gaps/Issues/Bug items in torchchatThese are known Gaps/Issues/Bug items in torchchat

MPS/MetalIssues related to Metal MPS/env set upIssues related to Metal MPS/env set up

performance

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1444 In pytorch/torchchat;

· swolchok opened on Dec 27, 2024

[KNOWN BUG] Broken Support for TextOnly Models from torchtune

bugSomething isn't workingSomething isn't working

Known GapsThese are known Gaps/Issues/Bug items in torchchatThese are known Gaps/Issues/Bug items in torchchat

torchtuneIssue/PR related to torchtune componentsIssue/PR related to torchtune components

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1430 In pytorch/torchchat;

· Jack-Khuu opened on Dec 19, 2024

Misaligned AOTI input; potential perf gains by fixing?

actionableItems in the backlog waiting for an appropriate impl/fixItems in the backlog waiting for an appropriate impl/fix

bugSomething isn't workingSomething isn't working

Compile / AOTIIssues related to AOT Inductor and torch compileIssues related to AOT Inductor and torch compile

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1424 In pytorch/torchchat;

· Jack-Khuu opened on Dec 14, 2024

OOM Llama 3.2-11B on RTX 4090

bugSomething isn't workingSomething isn't working

performance

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1341 In pytorch/torchchat;

· byjlw opened on Nov 2, 2024

False Positive on CI tests: test-readme

actionableItems in the backlog waiting for an appropriate impl/fixItems in the backlog waiting for an appropriate impl/fix

bugSomething isn't workingSomething isn't working

CI InfraIssues related to CI infrastructure and setupIssues related to CI infrastructure and setup

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1315 In pytorch/torchchat;

· Jack-Khuu opened on Oct 18, 2024

Llama 3.2 11B Currently Only Supports Single Image

enhancementNew feature or requestNew feature or request

Known GapsThese are known Gaps/Issues/Bug items in torchchatThese are known Gaps/Issues/Bug items in torchchat

Llama 3.2- MultimodalIssues related to Multimodal of Llama3.2Issues related to Multimodal of Llama3.2

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1281 In pytorch/torchchat;

· Jack-Khuu opened on Oct 8, 2024

GeneratorArgs.is_torchtune_model is a misnomer

actionableItems in the backlog waiting for an appropriate impl/fixItems in the backlog waiting for an appropriate impl/fix

good first issueGood for newcomersGood for newcomers

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1273 In pytorch/torchchat;

· Jack-Khuu opened on Oct 5, 2024

convert_hf_checkpoint only relies on model_name to resolve TransformerArgs

actionableItems in the backlog waiting for an appropriate impl/fixItems in the backlog waiting for an appropriate impl/fix

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1179 In pytorch/torchchat;

· Jack-Khuu opened on Sep 23, 2024

Slimming down torchchat: Replace replace_attention_with_custom_sdpa_attention() with ET's implementation

enhancementNew feature or requestNew feature or request

ExecuTorchIssues related to ExecuTorch installation, export, or build. Mobile uses separate tagsIssues related to ExecuTorch installation, export, or build. Mobile uses separate tags

good first issueGood for newcomersGood for newcomers

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#1058 In pytorch/torchchat;

· Jack-Khuu opened on Aug 23, 2024

Android App Should Crash gracefully when the tokenizer in the aar doesn't match the model

actionableItems in the backlog waiting for an appropriate impl/fixItems in the backlog waiting for an appropriate impl/fix

ExecuTorchIssues related to ExecuTorch installation, export, or build. Mobile uses separate tagsIssues related to ExecuTorch installation, export, or build. Mobile uses separate tags

Mobile - AndroidIssues Related to the Android WorkflowIssues Related to the Android Workflow

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#978 In pytorch/torchchat;

· Jack-Khuu opened on Jul 30, 2024

Improve the scope of Model Evaluation to AOTI and ET

enhancementNew feature or requestNew feature or request

Evaluation/BenchmarkingIssues Related to Evaluation and BenchmarkingIssues Related to Evaluation and Benchmarking

Known GapsThese are known Gaps/Issues/Bug items in torchchatThese are known Gaps/Issues/Bug items in torchchat

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#938 In pytorch/torchchat;

· Jack-Khuu opened on Jul 22, 2024