-
Notifications
You must be signed in to change notification settings - Fork 381
Pull requests: meta-recsys/generative-recommenders
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Set This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
separate_epilogue_store in persistent TMA addmm
CLA Signed
#531
opened May 28, 2026 by
njriasan
Contributor
Loading…
Fix AMD Triton ttgir compilation failure in concat/split_2D_jagged kernels
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#529
opened May 27, 2026 by
yuhuishi-convect
Loading…
Remove @triton_cc decorator on _addmm_fwd
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#528
opened May 27, 2026 by
cp2923
Contributor
Loading…
Add EPILOGUE_SUBTILE=8 and generalize TritonBench addmm kernel
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#527
opened May 22, 2026 by
jananisriram
Contributor
Loading…
Register more kernels as custom op
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#526
opened May 14, 2026 by
ruochen99
Contributor
Loading…
SymInt and FakeTensor tracing compatibility
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#525
opened May 14, 2026 by
ruochen99
Contributor
Loading…
Add AOT-T HSTU C++ runner code path
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#524
opened May 13, 2026 by
zoranzhao
Contributor
Loading…
add standalone triton_aot utils for dlrmv3 Triton AOT deployment
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#523
opened May 13, 2026 by
zoranzhao
Contributor
Loading…
TorchScript HSTU sparse + dense for C++ deployment
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#522
opened May 5, 2026 by
zoranzhao
Contributor
Loading…
Make HSTU Triton attention TLX path safe under enable_tma=False
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#521
opened May 5, 2026 by
zoranzhao
Contributor
Loading…
Wrap FB-only library loads and runtime imports with try/except
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#520
opened May 3, 2026 by
ruochen99
Contributor
Loading…
Enable Pyrefly in fbcode/generative_recommenders
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#519
opened May 1, 2026 by
maggiemoss
Loading…
Optimize rms_norm kernel for MI300X (HIP/AMD) (also benefits hammer wrapper)
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#518
opened Apr 29, 2026 by
mrmiywj
Contributor
Loading…
Fix non-contiguous tensor in helion split_2d_jagged backward
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#517
opened Apr 24, 2026 by
ruochen99
Contributor
Loading…
Fix redundant ctx.saved_tensors access in HSTU self attention
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#516
opened Apr 22, 2026 by
rguo-aws
Loading…
Rewrite PyTorch jagged ops to make them tracable
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#514
opened Apr 13, 2026 by
ruochen99
Contributor
Loading…
Make generative_recommenders ops compatible with make_fx tracing
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#513
opened Apr 13, 2026 by
LinjianMa
Contributor
Loading…
Add Meta AutoWS support for triton_addmm
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#512
opened Apr 10, 2026 by
njriasan
Contributor
Loading…
Rewrite PyTorch jagged ops to make them tracable
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
Reverted
#511
opened Apr 10, 2026 by
ruochen99
Contributor
Loading…
Support Pipelining TMA store and Data Partitioning with addmm
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#510
opened Apr 9, 2026 by
njriasan
Contributor
Loading…
Add __init__.py files for generative_recommenders package
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#509
opened Apr 9, 2026 by
ifding
Contributor
Loading…
Enable autotune for SHARDS_PER_SM on the bwd path
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#507
opened Apr 9, 2026 by
rguo-aws
Loading…
fix fp8 addmm when silu_u is False.
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#506
opened Apr 7, 2026 by
yaoyj11
Contributor
Loading…
Back out D97674495: keep contextual seq len in blackwell cutlass hstu attn kernel
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#505
opened Apr 6, 2026 by
yaoyj11
Contributor
Loading…
Fix AOT-T kernel selection for layer_norm ops, handle 3D dense jagged inputs, and add allow_tf32 to addmm
CLA Signed
This label is managed by the Meta Open Source bot.
fb-exported
meta-exported
#504
opened Apr 3, 2026 by
lurunming
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.