Wei-Cheng (Wayne) Chiu
is
Contributing...
@waynehacking8's recent pull requests on GitHub.
kernel queue
live
prs
76
completed
17
in-flight
59
//
making every FLOP count
Newset
Register wayne.is-a.dev
is-a-dev
/
register
10.7K
#44117
11 hours ago
[JIT] Fuse FP8 KV-cache quantization into the store kernel
sgl-project
/
sglang
30.5K
#31652
yesterday
fix(planner): resolve aggregate (mode=agg) DGDs via agg_worker_k8s_name
ai-dynamo
/
dynamo
7.5K
#11578
last week
[#16196][fix] Support MIXED_PRECISION experts-only NVFP4 DeepSeek-V4 checkpoints
NVIDIA
/
TensorRT-LLM
14.2K
#16276
last week
[Quantization] Fix misleading NVFP4 Marlin linear warning on FP4-native GPUs
vllm-project
/
vllm
86.6K
#48199
last week
[Bugfix] Handle E8M0 block scales in CUTLASS and Triton FP8 linear kernels
vllm-project
/
vllm
86.6K
#47988
2 weeks ago
fix(attention): correct SM120 NVFP4 qk_correction layout, row-sum reduction, and lse
flashinfer-ai
/
flashinfer
6K
#3838
2 weeks ago
[None][fix] Exclude SM12x from bf16x2 FP8 cvt specialization in moe_a2a
NVIDIA
/
TensorRT-LLM
14.2K
#15934
2 weeks ago
fix(b12x): bucket MoE workspace allocation to stop kernel-cache thrash
flashinfer-ai
/
flashinfer
6K
#3836
2 weeks ago
fix(sparse-mla-sm120): add TOPK=256 DSV4 prefill instantiation
flashinfer-ai
/
flashinfer
6K
#3834
2 weeks ago
feat(vllm): support NixlPushConnector in PD coordination and NIXL detection
ai-dynamo
/
dynamo
7.5K
#11219
2 weeks ago
fix(config): read num_nextn_predict_layers from the outer hf_config as fallback
sgl-project
/
sglang
30.5K
#30092
2 weeks ago
test: seed test_mm_fp4 inputs to fix flaky cos_sim threshold failures
flashinfer-ai
/
flashinfer
6K
#3829
2 weeks ago
[MoE/b12x] Auto-select FLASHINFER_B12X NVFP4 MoE backend on SM120, reject EP deployments
vllm-project
/
vllm
86.6K
#47577
2 weeks ago
test(comm): order rank streams after input fills in dcp_alltoall tests
flashinfer-ai
/
flashinfer
6K
#3815
2 weeks ago
fix(common): make multimodal media loaders importable without torch
ai-dynamo
/
dynamo
7.5K
#11179
2 weeks ago
test(sgl-kernel): gate arch-specific tests that fail on SM120 (253 expected-unsupported failures)
sgl-project
/
sglang
30.5K
#29902
2 weeks ago
fix(config): derive scaling in the KimiVL, MiniCPM3, and DeepseekVL2 MLA branches
sgl-project
/
sglang
30.5K
#29896
2 weeks ago
fix(config): treat v_head_dim=0 as unset when deriving model shapes (deepseek-vl2-tiny warmup crash)
sgl-project
/
sglang
30.5K
#29890
2 weeks ago
[SM120] Emit grouped NVFP4 (ue4m3) block-scaled GEMM in cutlass_library
NVIDIA
/
cutlass
10.1K
#3346
4 weeks ago
[Kernel][SM120] NVFP4 grouped MoE: pingpong schedule at large per-exp…
vllm-project
/
vllm
86.6K
#46481
4 weeks ago
fix(trace): handle combined paged KV cache tensor in _get_tensor (#3685)
flashinfer-ai
/
flashinfer
6K
#3690
4 weeks ago
fix(server_args): gate trtllm_mla prefill (not decode) off SM120/SM121
sgl-project
/
sglang
30.5K
#28837
4 weeks ago
[Bugfix] Forward upstream error message in Anthropic streaming converter (#46028)
vllm-project
/
vllm
86.6K
#46162
last month
fix(logging): allow exporting OTLP traces without logs
ai-dynamo
/
dynamo
7.5K
#10806
last month
[CuTeDSL] Fix make_ptr ZeroDivisionError on sub-byte dtypes
NVIDIA
/
cutlass
10.1K
#3334
last month
[CuTeDSL] Fix cute.is_major crash on List[int] mode
NVIDIA
/
cutlass
10.1K
#3333
last month
[Bugfix] Fix ImportError in get_chat_template for builtin chat-template names
InternLM
/
lmdeploy
8K
#4690
last month
fix(moe): don't mark trtllm-gen FP4 MoE backend supported on SM120/SM121
flashinfer-ai
/
flashinfer
6K
#3670
last month
fix(mla): don't recommend backend='cutlass' on SM120/SM121 MLA fallback
flashinfer-ai
/
flashinfer
6K
#3669
last month
fix(mamba): reject SM120/SM121 in SSDCombined with a clear error
flashinfer-ai
/
flashinfer
6K
#3668
last month
[Bugfix] Fix double-counted max_q_seqlen in decode delta kv_seqlens
InternLM
/
lmdeploy
8K
#4685
last month
[Bugfix] Fix InternVL/InternVL3 LoRA loading TypeError in adapter fallback
InternLM
/
lmdeploy
8K
#4684
last month
[Bugfix] Clear reqs_status on async lookup timeout to prevent KeyError on recall
LMCache
/
LMCache
10.7K
#3722
last month
[Offload] Skip device-map entries with no local module in dispatch_with_map
vllm-project
/
compressed-tensors
302
#737
last month
[Bugfix] Don't fail model load for multimodal models with no language model (#45806)
vllm-project
/
vllm
86.6K
#45811
last month
[Bugfix] DiffusionGemma: only pop a request's logprobs when it commits (#45689)
vllm-project
/
vllm
86.6K
#45754
last month
fix(moe): return bf16 expert_weights from trtllm_fp4_block_scale_moe (#3595)
flashinfer-ai
/
flashinfer
6K
#3644
last month
[Fix] Update speculative draft weights on update_weights_from_distributed
sgl-project
/
sglang
30.5K
#28257
last month
fix(jit): quote -ccbin host compiler path so it works with spaces
flashinfer-ai
/
flashinfer
6K
#3637
last month
fix(jit): cap ninja parallelism by available memory to avoid OOM (#3634)
flashinfer-ai
/
flashinfer
6K
#3636
last month
fix: correct MHA PP KV-pointer slicing for uneven full-attn splits (#27740)
sgl-project
/
sglang
30.5K
#28229
last month
[Bugfix] Bounds-check moe_permute reverse-map write (#45492)
vllm-project
/
vllm
86.6K
#45530
last month
[Kernel] Add SM120/SM121 dispatch for int8_scaled_mm
sgl-project
/
sglang
30.5K
#28137
last month
[PT2E][X86] Extend reused-module fallback to nn.Conv2d (follow-up to #4480)
pytorch
/
ao
2.9K
#4482
last month
feat(moe): add SM120/SM121 dispatch for fp8_blockwise_scaled_grouped_mm
sgl-project
/
sglang
30.5K
#28125
last month
fix: give concurrent streams disjoint row_states buffers in multi-CTA top-k
flashinfer-ai
/
flashinfer
6K
#3625
last month
[Bugfix] Reject structured outputs for diffusion decoders with a clear error
vllm-project
/
vllm
86.6K
#45468
last month
[PT2E][X86] Fall back to plain linear annotation for reused nn.Linear with fusable post op
pytorch
/
ao
2.9K
#4480
last month
[Bugfix] Return the tokenizer from maybe_make_thread_pool so it survives pickling
vllm-project
/
vllm
86.6K
#45460
last month
Add tuned fused-MoE Triton config for Gemma-4 FP8 on RTX PRO 6000 Blackwell (SM120)
sgl-project
/
sglang
30.5K
#28077
last month
fix(topk): eliminate multi-CTA radix top-k stream hangs on SM120/SM121
flashinfer-ai
/
flashinfer
6K
#3615
last month
[BugFix] Don't strand parked async-KV-load requests behind an unschedulable queue head
vllm-project
/
vllm
86.6K
#45406
last month
Clamp default fused-MoE Triton configs to the device shared memory limit (SM120/SM121)
sgl-project
/
sglang
30.5K
#28038
last month
[Kernel] Resurrect rotted gather-cache OOB test + defensive block-table bound-checks
vllm-project
/
vllm
86.6K
#45393
last month
fix: make the cutlass MoE gemm profiler MXFP8-aware (autotune crash on MXFP8xMXFP8)
flashinfer-ai
/
flashinfer
6K
#3614
last month
[Bugfix] Fix int32 overflow in concat_mla_q flat warp index
vllm-project
/
vllm
86.6K
#45384
last month
test: fix CUDA OOM in batch-prefill custom-mask test on 24GB CI GPUs
flashinfer-ai
/
flashinfer
6K
#3609
last month
Fix BF16 RL weight update with non-routed flashinfer_trtllm MoE backend
sgl-project
/
sglang
30.5K
#28015
last month
[Bugfix] Set type/role explicitly in streaming message_start event
vllm-project
/
vllm
86.6K
#45376
last month
fix: apply rope overrides to text_config for VLM-format models
sgl-project
/
sglang
30.5K
#27983
last month
[Bugfix] Make block-fp8 fused MoE test tolerance noise-aware
vllm-project
/
vllm
86.6K
#45334
last month
[Bugfix] Skip Ray actor kill after the driver has disconnected
vllm-project
/
vllm
86.6K
#45328
last month
test: bound peak memory in mxfp4 groupwise group GEMM test (32GB OOM)
flashinfer-ai
/
flashinfer
6K
#3594
last month
[Bugfix] Abort requests that arrive while the engine is asleep
vllm-project
/
vllm
86.6K
#45326
last month
test: cover diffkv (qk192/v128) + attention-sink triton attention kernels
sgl-project
/
sglang
30.5K
#27963
last month
fix(attention): sm120 shared-memory-safe block size in context_attention_fwd
sgl-project
/
sglang
30.5K
#27962
last month
[Bugfix] Reject NVFP4 MoE checkpoints with missing per-expert scales
vllm-project
/
vllm
86.6K
#45320
last month
fix(moe): thread calibrated NVFP4 global scales through the unified MoE API
flashinfer-ai
/
flashinfer
6K
#3592
last month
test(moe): align EP-offset tests and docs with global expert ids (gh #3547)
flashinfer-ai
/
flashinfer
6K
#3591
last month
fix(kv-router): cancel in-flight recovery when worker is removed
ai-dynamo
/
dynamo
7.5K
#10616
last month
Fix MiMo-V2 expert routing on TopK-bypassing MoE backends
sgl-project
/
sglang
30.5K
#27916
last month
test: cover XQA decode head_dim=256 with small page sizes (#2638)
flashinfer-ai
/
flashinfer
6K
#3577
last month
feat: add RTX PRO 6000 Blackwell GPU SKU
ai-dynamo
/
dynamo
7.5K
#10579
last month
[#14146][bugfix] Fix CuteDSL MoE ghost-token and global-index bugs
NVIDIA
/
TensorRT-LLM
14.2K
#15155
last month
[None][bugfix] Guard e_score_correction_bias load for DeepSeek-V2-Lite
NVIDIA
/
TensorRT-LLM
14.2K
#15154
last month
Source code
Deploy with Vercel
