Baizhou Zhang 39008955ff Revert "[AMD][MORI] Fix MTP crash with FP4/FP8 dispatch and add NEXTN dispatch env vars." (#20602)

2026-03-14 12:12:42 -07:00

..

attention

[Bug-fix] Fix gpu fault when run the test with dp-attention-enabled and max-concurrency is over 256 (#20399)

2026-03-12 02:32:03 -07:00

deep_gemm_wrapper

[DSv32] [GLM5] Improve Model Quality by Avoiding FP32 Precision Loss in weights_proj (#19041)

2026-02-22 16:20:51 +08:00

moe

Revert "[AMD][MORI] Fix MTP crash with FP4/FP8 dispatch and add NEXTN dispatch env vars." (#20602)

2026-03-14 12:12:42 -07:00

quantization

[Model] Support Nemotron 3 Super NVFP4 (#20407)

2026-03-14 00:56:26 -07:00

rotary_embedding

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

utils

Support return_logprob for spec v2 (overlap safe) (#19801)

2026-03-10 15:38:27 -07:00

activation.py

…

amx_utils.py

…

communicator_nsa_cp.py

refactor context parallel state (#17213)

2026-02-13 23:18:17 +08:00

communicator.py

[Fix] Add fallback for flashinfer allreduce fusion (#20384)

2026-03-13 01:24:55 -07:00

dp_attention.py

Fix comment for tp_rank calculation in dp_attention (#19306)

2026-02-25 15:19:10 +08:00

elementwise.py

…

flashinfer_comm_fusion.py

[Fix] Add fallback for flashinfer allreduce fusion (#20384)

2026-03-13 01:24:55 -07:00

int4fp8_utils.py

…

layernorm.py

[Fix] Add fallback for flashinfer allreduce fusion (#20384)

2026-03-13 01:24:55 -07:00

linear.py

[Feature][NPU]: add runtime support for AutoRound quantized models (#16699)

2026-03-07 18:03:55 +03:00

logits_processor.py

Support return_logprob for spec v2 (overlap safe) (#19801)

2026-03-10 15:38:27 -07:00

model_parallel.py

…

modelopt_utils.py

…

multimodal.py

…

n_gram_embedding.py

[1/n jit_kernel restruct] unify cache usage and clean up naming in ngram_embedding (#20244)

2026-03-10 15:53:43 +08:00

parameter.py

[quant] fix fp32 downcasting (#19844)

2026-03-05 17:54:59 +08:00

pooler.py

…

radix_attention.py

update pre-commit config (#18860)

2026-02-16 00:18:31 +08:00

radix_linear_attention.py

[PCG] support piecewise cuda graph for kimi-linear model (#18849)

2026-02-17 23:31:12 +08:00

rocm_linear_utils.py

…

sampler.py

use flashinfer.sampling (#18696)

2026-02-26 10:02:38 +08:00

sparse_pooler.py

…

torchao_utils.py

…

vocab_parallel_embedding.py

…