Ifu dev 20260803 v2.18 merge - #720

Merged

matthiasdiener merged 49 commits into

devfrom

IFU-dev-20260803-v2.18-merge

Aug 28, 2026

Commits on Jun 23, 2026

Changed VERSION to 2.18.0.dev0 (#3140)
KshitijLakhani
authored
43206d9
[Common] Add dense router output for fused router (#3129)
harryzhou2000
authored
77054fa
Update FE to 1.25 (#3139)
cyanguwa
authored
cd20e9d

Commits on Jun 24, 2026

[Common][PyTorch]Current Scaling Group Quantization + Enabling Varying Last/Both Dims in Group Quantize (#3114)

authored
ab66858
[PyTorch] Enable Weight Preswizzling during Quantization in TE Modules (#3093)

authored
42cb81b

Commits on Jun 25, 2026

Commits on Jun 26, 2026

Add L2 score mod distributed attention shape (#3147)
vcherepanov-nv
authored
2c7d5a0

Commits on Jun 27, 2026

Commits on Jun 29, 2026

Commits on Jun 30, 2026

Commits on Jul 1, 2026

Commits on Jul 2, 2026

docs: document attention backend selection (#3142)
sbhavani
authored
f721112

Commits on Jul 3, 2026

Commits on Jul 6, 2026

Commits on Jul 7, 2026

Commits on Jul 8, 2026

[ALL] Support int32 topk_idx in TE EP (#3185)
phu0ngng
authored
20a4912
docs(readme): update latest news (#3121)
sbhavani
authored
8b99682

Commits on Jul 9, 2026

Commits on Jul 10, 2026

Remove unused fused attention workspace optimization control (#3180)

ptrendx
and
pre-commit-ci[bot]
authored
7dbdf7b
Enable NVFP4 RHT amax for grouped SReLU MLP (#3133)

authored
dfa0744
TE EP integration to MoEBlock (#3116)

tdophung
and
pre-commit-ci[bot]
authored
b782ea6
[PyTorch] Fix GIL/refcount abort in Comm+GEMM overlap and NCCL-EP init bindings (#3203)
denera
authored
6377ca1

Commits on Jul 14, 2026

Commits on Jul 15, 2026

Commits on Jul 17, 2026

[PyTorch] Add optional caller-provided output/grad-input buffers to GroupedLinear and fused grouped MLP (#3161)

authored
68493d2

Commits on Jul 20, 2026

Fix FusedAdam empty tensor handling (#3212)

wujingyue
and
vthumbe1503
authored
868d8d9

Commits on Aug 27, 2026