GitHub · Where software is built

Labels Milestones New issue

Switch to pytorch-style CI docker image naming

#1399

· jithunnair-amd opened

on Jul 15, 2025

Circular imports

#1383

· ebsmothers opened

on Jul 12, 2025

Llama4 conversion scripts error on aux-loss-free params

#1379

· jquesnelle opened

on Jul 10, 2025

Puzzling collectives in TP ( SP to be exact)

#1369

· githubsgi opened

on Jul 7, 2025

Support more models such as Qwen

#1368

· fzyzcjy opened

on Jul 4, 2025

Data loader's state_dict is being lost between being saved and loaded when the dataset loops

#1357

· wesleytruong opened

on Jun 30, 2025

Llama4 TP bug: DTensor local tensor dtype does not match DTensorSpec tensor meta dtype, causing meta registration error

#1355

· danielvegamyhre opened

on Jun 28, 2025

Issue reproducing Float8 performance benchmark

#1344

· xmfan opened

on Jun 26, 2025

[Low-bit Optimizers] Do torchtitan plan to integrate AdamW8bit or AdamWFP8 from TorchAO

#1334

· haochengxi opened

on Jun 24, 2025

OOM recovery under multi-node FSDP/HSDP

#1329

· vadimkantorov opened

on Jun 23, 2025

Why <code>preserve_rng_state=False</code> in activation checkpointing

#1323

· awgu opened

on Jun 20, 2025

How to adapt HuggingFace or other models for TorchTitan

#1322

· githubsgi opened

on Jun 20, 2025