repo stringclasses 454
values | file_path stringlengths 5 201 | extension stringclasses 1
value | content stringlengths 8 509k | num_lines int64 3 16.9k | size_bytes int64 8 511k |
|---|---|---|---|---|---|
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_mxfp4_w4a8.py | .py | """ModelSlim W4A8_MXFP scheme for pre-quantized weight inference on Ascend NPU (SRT).
The msmodelslim ``W4A8_MXFP`` checkpoint stores weights as **packed FP4**:
weight: uint8 (pack_fp4_to_uint8), shape [out, in//2], group_size=32
weight_scale: uint8 (UE8M0, +127 biased), shape [out, in//32]
(verified... | 108 | 4,499 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_scheme.py | .py | # Adapted from https://github.com/vllm-project/vllm/tree/main/vllm/model_executor/layers/quantization/compressed_tensors
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from abc import abstractmethod
from typing import Optional
import torch
from sglang.srt.l... | 75 | 2,272 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_w8a8_int8.py | .py | # Adapted from https://github.com/vllm-project/vllm/tree/main/vllm/model_executor/layers/quantization/compressed_tensors
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from typing import Dict, List, Optional
import torch
from sglang.srt.hardware_backend.npu... | 119 | 4,171 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_w4a4_int4_moe.py | .py | from __future__ import annotations
import logging
from typing import Any, Dict
import torch
from sglang.srt.environ import envs
from sglang.srt.hardware_backend.npu.quantization.moe_methods import (
NPUW4A4Int4MoEMethod,
)
from sglang.srt.layers.quantization.modelslim.schemes import ModelSlimMoEScheme
from sglan... | 138 | 4,584 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_mxfp8.py | .py | """ModelSlim MXFP8 scheme for pre-quantized weight inference on Ascend NPU (SRT).
Loads weights pre-quantized by msmodelslim (float8_e4m3fn weights,
uint8 scales) and runs MXFP8 matmul at inference.
Following the modelslim-scheme convention (see ModelSlimW8A8Int8), this scheme
owns only the hardware-agnostic weight c... | 90 | 3,283 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_mxfp4.py | .py | """ModelSlim W4A4_MXFP4 scheme for pre-quantized weight inference on Ascend NPU (SRT).
The msmodelslim ``W4A4_MXFP4`` checkpoint stores weights as **packed FP4**:
weight: uint8 shape [out, in//2] (two FP4 values per byte)
weight_scale: uint8 (UE8M0) shape [out, in//32] (block scales, group_s... | 99 | 3,697 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_w4a4_int4.py | .py | # Adapted from https://github.com/vllm-project/vllm/tree/main/vllm/model_executor/layers/quantization/compressed_tensors
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from typing import Any, Dict, List, Optional
import torch
from sglang.srt.environ import ... | 114 | 4,306 |
sglang | python/sglang/srt/layers/quantization/modelslim/schemes/modelslim_mxfp8_moe.py | .py | """ModelSlim MXFP8 offline scheme for MoE layers on Ascend NPU (SRT).
Loads weights pre-quantised by msmodelslim: float8_e4m3fn weights + uint8 block
scales (block_size=32). The layout transform and the forward pass are delegated
to ``NPUMXFP8MoEMethod`` -- the same kernel the online MXFP8 MoE path uses.
"""
from __f... | 115 | 4,261 |
sglang | python/sglang/srt/layers/cp/utils.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 343 | 12,012 |
sglang | python/sglang/srt/layers/cp/__init__.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 56 | 1,678 |
sglang | python/sglang/srt/layers/cp/bcg.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 310 | 12,103 |
sglang | python/sglang/srt/layers/cp/padding.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 64 | 2,409 |
sglang | python/sglang/srt/layers/cp/cp_decode_attn_tp.py | .py | """CP Decode Attention TP context.
When CP (Context Parallel) mode sets tp_size=1 (repeat weights), decode can
partition attention weights across CP ranks matching normal TP behavior.
"""
from __future__ import annotations
import logging
from contextlib import contextmanager
from typing import TYPE_CHECKING, Dict, L... | 214 | 8,551 |
sglang | python/sglang/srt/layers/cp/zigzag.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 479 | 18,496 |
sglang | python/sglang/srt/layers/cp/base.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 308 | 9,683 |
sglang | python/sglang/srt/layers/cp/interleave.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 273 | 10,207 |
sglang | python/sglang/srt/layers/rotary_embedding/utils.py | .py | """Primitive rotary embedding ops: _rotate_neox, _rotate_gptj, _apply_rotary_emb,
apply_rotary_pos_emb variants."""
from __future__ import annotations
from typing import Tuple
import torch
from sglang.srt.utils import cpu_has_amx_support, get_compiler_backend, is_cpu, is_npu
_is_npu = is_npu()
_is_cpu = is_cpu()
_... | 147 | 4,113 |
sglang | python/sglang/srt/layers/rotary_embedding/rope_variant.py | .py | """RoPE scaling variants: Phi3LongRoPE, FourierRoPE, DeepseekScaling, Llama3,
Llama4Vision, DynamicNTK, DynamicNTKAlpha, DualChunkRotaryEmbedding."""
from __future__ import annotations
import math
from typing import List, Optional, Tuple, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from... | 938 | 34,340 |
sglang | python/sglang/srt/layers/rotary_embedding/yarn.py | .py | """YaRNScalingRotaryEmbedding + YaRN helper functions."""
from __future__ import annotations
import math
from typing import Tuple
import torch
from sglang.srt.layers.rotary_embedding.base import RotaryEmbedding
# Inverse dim formula to find dim based on number of rotations
def yarn_find_correction_dim(
num_ro... | 147 | 4,678 |
sglang | python/sglang/srt/layers/rotary_embedding/__init__.py | .py | # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# Adapted from https://raw.githubusercontent.com/vllm-project/vllm/refs/tags/v0.6.6.post1/vllm/model_executor/layers/rotary_embedding.py
"""Rotary Positional Embeddings - public API (drop-in replacement for rotary... | 34 | 1,218 |
sglang | python/sglang/srt/layers/rotary_embedding/factory.py | .py | """Factory functions: get_rope, get_rope_cpu, get_rope_wrapper."""
from __future__ import annotations
import logging
from typing import Any, Dict, Optional, Tuple
import torch
from sglang.srt.layers.rotary_embedding.base import (
LinearScalingRotaryEmbedding,
RotaryEmbedding,
)
from sglang.srt.layers.rotary... | 493 | 16,095 |
sglang | python/sglang/srt/layers/rotary_embedding/mrope_rope_index.py | .py | """get_rope_index implementations for Qwen2-VL/Qwen3-VL, Qwen3-Omni, GLM4V, Ernie4.5."""
from __future__ import annotations
import itertools
from typing import Any, List, Optional, Tuple, Union
import torch
def _get_feat_extract_output_lengths(input_lengths):
"""
Computes the output length of the convoluti... | 830 | 34,309 |
sglang | python/sglang/srt/layers/rotary_embedding/mrope.py | .py | """MRotaryEmbedding, YaRNScalingMRotaryEmbedding, Ernie4_5_VLRotaryEmbedding,
apply_interleaved_rope for multimodal RoPE."""
from __future__ import annotations
from typing import List, Optional, Tuple
import torch
from sglang.kernels.ops.attention.rotary_triton import (
triton_ernie45_rope_fused_inplace,
tr... | 615 | 21,836 |
sglang | python/sglang/srt/layers/rotary_embedding/base.py | .py | """RotaryEmbedding base class + LinearScalingRotaryEmbedding."""
from __future__ import annotations
import logging
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, Union
import torch
from sglang.kernels.fused_op import BaseFusedOp
from sglang.srt.environ import envs
from sglang.srt.layers.rotary_embed... | 571 | 21,501 |
sglang | python/sglang/srt/layers/moe/cutlass_moe.py | .py | """CUTLASS based Fused MoE kernels."""
from typing import Optional, Tuple
import torch
from sglang.srt.utils import is_cuda, is_sm90_supported, is_sm100_supported
_is_cuda = is_cuda()
if _is_cuda:
from sgl_kernel import (
apply_shuffle_mul_sum,
es_fp8_blockwise_scaled_grouped_mm,
es_sm10... | 337 | 13,482 |
sglang | python/sglang/srt/layers/moe/utils.py | .py | from __future__ import annotations
import logging
import os
from contextlib import contextmanager
from enum import Enum, IntEnum
from typing import TYPE_CHECKING
import torch
from sglang.srt.environ import envs
from sglang.srt.layers.dp_attention import (
is_dp_attention_enabled,
)
from sglang.srt.runtime_contex... | 766 | 27,838 |
sglang | python/sglang/srt/layers/moe/flashinfer_trtllm_moe.py | .py | from typing import Optional
import torch
from sglang.srt.utils.custom_op import register_custom_op
def _fake_fp8_block_scale_moe_out(
routing_logits: torch.Tensor,
routing_bias: Optional[torch.Tensor],
hidden_states: torch.Tensor,
hidden_states_scale: torch.Tensor,
gemm1_weights: torch.Tensor,
... | 324 | 11,001 |
sglang | python/sglang/srt/layers/moe/flashinfer_cutedsl_moe.py | .py | from typing import Optional
import torch
from flashinfer import (
scaled_fp4_grouped_quantize,
silu_and_mul_scaled_nvfp4_experts_quantize,
)
from flashinfer.cute_dsl.blockscaled_gemm import grouped_gemm_nt_masked
def get_cute_dtype(input: torch.Tensor) -> str:
if input.dtype == torch.bfloat16:
re... | 210 | 7,387 |
sglang | python/sglang/srt/layers/moe/mega_moe.py | .py | # Copyright 2023-2024 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 384 | 12,768 |
sglang | python/sglang/srt/layers/moe/__init__.py | .py | from sglang.srt.layers.moe.moe_runner import MoeRunner, MoeRunnerConfig
from sglang.srt.layers.moe.utils import (
DeepEPMode,
MoeA2ABackend,
MoeRunnerBackend,
get_deepep_config,
get_deepep_mode,
get_moe_a2a_backend,
get_moe_runner_backend,
get_tbo_token_distribution_threshold,
initia... | 37 | 1,007 |
sglang | python/sglang/srt/layers/moe/topk.py | .py | # Copyright 2024 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, so... | 2,375 | 93,411 |
sglang | python/sglang/srt/layers/moe/hash_topk.py | .py | from __future__ import annotations
import logging
from typing import Optional, Tuple
import torch
from torch import nn
from sglang.srt.environ import envs
from sglang.srt.eplb.expert_distribution import (
get_global_expert_distribution_recorder,
)
from sglang.srt.eplb.expert_location_dispatch import (
Expert... | 276 | 10,780 |
sglang | python/sglang/srt/layers/moe/fused_moe_native.py | .py | """
Torch-native implementation for FusedMoE. This is used for torch.compile.
It is based on https://github.com/pytorch-labs/gpt-fast/blob/32971d3129541c5bfb4f715abc33d1c5f408d204/mixtral-moe/model.py#L204
"""
import torch
from torch.nn import functional as F
from sglang.srt.layers.activation import GeluAndMul, SiluA... | 165 | 5,738 |
sglang | python/sglang/srt/layers/moe/cutlass_moe_params.py | .py | from dataclasses import dataclass
from enum import Enum, auto
from typing import Optional
import torch
class CutlassMoEType(Enum):
"""
Enum for the different types of cutlass moe operations
that are currently supported in SGLang.
"""
BlockscaledFP8 = auto()
BlockscaledFP4 = auto()
@datacla... | 188 | 7,226 |
sglang | python/sglang/srt/layers/moe/cutlass_w4a8_moe.py | .py | # SPDX-License-Identifier: Apache-2.0
"""Cutlass W4A8 MoE kernel."""
from typing import Optional
import torch
from sglang.srt.runtime_context import get_parallel
from sglang.srt.utils import is_cuda, is_cuda_alike
_is_cuda = is_cuda()
_is_cuda_alike = is_cuda_alike()
if _is_cuda_alike:
from sgl_kernel import (... | 578 | 20,500 |
sglang | python/sglang/srt/layers/moe/kt_ep_wrapper.py | .py | # SPDX-License-Identifier: Apache-2.0
"""
KT Expert Parallelism Wrapper for MoE layers.
This module provides a generic wrapper that enables CPU-GPU expert parallelism
for any MoE quantization method. It coordinates parallel execution of GPU experts
(using any quantization method) and CPU experts (using AMX/AVX instruc... | 394 | 14,240 |
sglang | python/sglang/srt/layers/moe/route_quant_handoff.py | .py | """Attempt-and-verify handoff for the fused K3 MoE-front prep launch.
At decode batch sizes the chain between the K3 fused-front GEMM and the
trtllm-gen SiTU MoE op is three tiny back-to-back kernels on the critical
path — route_radix (top-16 of 896), the triton ``(id << 16) | bf16(weight)``
pack, and per_token_group_... | 131 | 4,575 |
sglang | python/sglang/srt/layers/moe/waterfill.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 293 | 10,932 |
sglang | python/sglang/srt/layers/moe/mega_moe_sm90.py | .py | # Copyright 2023-2024 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 180 | 5,891 |
sglang | python/sglang/srt/layers/moe/fused_moe_triton/layer.py | .py | # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# Adapted from https://github.com/vllm-project/vllm/blob/a6221a144af772fd1a68fe7e627935dc53e81738/vllm/model_executor/layers/fused_moe/layer.py
import logging
from enum import Enum
from functools import cached_pr... | 1,743 | 67,984 |
sglang | python/sglang/srt/layers/moe/fused_moe_triton/__init__.py | .py | from sglang.srt.layers.moe.fused_moe_triton.layer import (
FusedMoE,
FusedMoeWeightScaleSupported,
)
from sglang.srt.layers.moe.moe_runner.triton_utils import (
fused_experts,
get_config,
get_config_file_name,
moe_align_block_size,
override_config,
try_get_optimal_moe_config,
)
__all__ ... | 24 | 530 |
sglang | python/sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py | .py | # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# Adapted from https://github.com/vllm-project/vllm/pull/18595/files#diff-f426a6de78c82ffec568eff6811bfbf0043dab5f87f1a8c0cffdbdcb8a81e035
from __future__ import annotations
from typing import TYPE_CHECKING, Opt... | 364 | 11,649 |
sglang | python/sglang/srt/layers/moe/fused_moe_triton/fused_marlin_moe.py | .py | from typing import Optional
import torch
import torch.nn.functional as F
import triton
import triton.language as tl
from sglang.srt.layers import zero_copy_context
from sglang.srt.utils import is_cuda
from sglang.srt.utils.custom_op import register_custom_op
_is_cuda = is_cuda()
if _is_cuda:
from sgl_kernel imp... | 429 | 14,599 |
sglang | python/sglang/srt/layers/moe/ep_moe/layer.py | .py | from __future__ import annotations
import logging
from typing import TYPE_CHECKING, Any, Dict, Optional
import torch
from sglang.kernels.ops.quantization.fp8_kernel import is_fp8_fnuz
from sglang.srt.environ import envs
from sglang.srt.layers import deep_gemm_wrapper
from sglang.srt.layers.dp_attention import (
... | 363 | 12,080 |
sglang | python/sglang/srt/layers/moe/dwdp/layout.py | .py | # Adapted from NVIDIA TensorRT-LLM (https://github.com/NVIDIA/TensorRT-LLM)
"""Expert ownership and page-aligned memory layout computation for DWDP."""
from __future__ import annotations
import math
from typing import Dict, List, Optional, Tuple
import torch
from sglang.srt.cuda_vmm_utils import align_down, align_u... | 288 | 8,769 |
sglang | python/sglang/srt/layers/moe/dwdp/transport.py | .py | # Adapted from NVIDIA TensorRT-LLM (https://github.com/NVIDIA/TensorRT-LLM)
"""Cross-rank expert weight handle exchange (FABRIC or POSIX fd) and peer view import."""
from __future__ import annotations
import logging
import os
from typing import Dict, List, Optional, Tuple
import torch
import torch.distributed as dis... | 234 | 8,084 |
sglang | python/sglang/srt/layers/moe/dwdp/page_pool.py | .py | # Adapted from NVIDIA TensorRT-LLM (https://github.com/NVIDIA/TensorRT-LLM)
"""Double-buffered pool of local VMM pages backing the remote regions of the composite VA."""
from __future__ import annotations
import logging
from typing import Dict, List, Optional
from cuda.bindings import driver as cuda
from sglang.srt... | 127 | 3,935 |
sglang | python/sglang/srt/layers/moe/dwdp/__init__.py | .py | """DWDP (Distributed Weight Data Parallelism): MoE prefill with tokens kept on-rank and peer expert weights prefetched via NVLink into a composite VMM address space."""
from sglang.srt.layers.moe.dwdp.dwdp_manager import DwdpManager
from sglang.srt.runtime_context import (
get_global_dwdp_manager,
set_global_d... | 14 | 431 |
sglang | python/sglang/srt/layers/moe/dwdp/weight_manager.py | .py | # Adapted from NVIDIA TensorRT-LLM (https://github.com/NVIDIA/TensorRT-LLM)
"""Double-buffered async prefetch of peer expert weights into the composite VA."""
from __future__ import annotations
import bisect
import logging
from typing import Dict, List, Optional, Tuple
import torch
from sglang.srt.layers.moe.dwdp.l... | 143 | 5,522 |
sglang | python/sglang/srt/layers/moe/dwdp/weight_buffer.py | .py | # Adapted from NVIDIA TensorRT-LLM (https://github.com/NVIDIA/TensorRT-LLM)
"""Composite VA presenting a contiguous full-expert weight tensor per (layer, weight)."""
from __future__ import annotations
import logging
from typing import Dict, List, Optional, Tuple
import torch
from sglang.srt.cuda_vmm_utils import (
... | 215 | 7,396 |
sglang | python/sglang/srt/layers/moe/dwdp/dwdp_manager.py | .py | """Global singleton orchestrating the DWDP lifecycle from setup(model) to cleanup()."""
from __future__ import annotations
import logging
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple
import torch
import torch.distributed as dist
from torch import nn
from sglang.srt.layers.moe.dwdp.layout import (
... | 212 | 7,983 |
sglang | python/sglang/srt/layers/moe/moe_runner/flashinfer_cutedsl.py | .py | from __future__ import annotations
import logging
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Optional
import torch
from sglang.srt.environ import envs
from sglang.srt.layers.moe.moe_runner.base import (
MoeQuantInfo,
MoeRunnerConfig,
register_fused_func,
)
from sglang.srt.mo... | 658 | 25,709 |
sglang | python/sglang/srt/layers/moe/moe_runner/marlin.py | .py | from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, Optional
import torch
import triton
import triton.language as tl
from sglang.srt.layers.moe.moe_runner.base import (
MoeQuantInfo,
MoeRunnerConfig,
RunnerInput,
RunnerOutput,
register_fused_func... | 219 | 7,703 |
sglang | python/sglang/srt/layers/moe/moe_runner/deep_gemm.py | .py | from __future__ import annotations
import logging
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, List, Optional, Tuple
import einops
import torch
import triton
import triton.language as tl
from sglang.kernels.ops.attention.dsv4 import silu_and_mul_masked_post_quant
from sglang.kernels.ops.q... | 1,451 | 52,700 |
sglang | python/sglang/srt/layers/moe/moe_runner/runner.py | .py | from __future__ import annotations
import logging
import os
from typing import TYPE_CHECKING, Any, Optional
from sglang.srt.layers.moe.moe_runner.base import (
FusedOpPool,
MoeRunnerConfig,
PermuteMethodPool,
)
from sglang.srt.layers.moe.moe_runner.deep_gemm import DeepGemmRunnerCore
from sglang.srt.layer... | 220 | 9,498 |
sglang | python/sglang/srt/layers/moe/moe_runner/__init__.py | .py | from sglang.srt.layers.moe.moe_runner.base import MoeRunnerConfig
from sglang.srt.layers.moe.moe_runner.runner import MoeRunner
__all__ = ["MoeRunnerConfig", "MoeRunner"]
| 5 | 172 |
sglang | python/sglang/srt/layers/moe/moe_runner/ascend.py | .py | """Ascend MoE runner backend with NPU‑specific ops."""
from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Optional
import torch
from sglang.srt.hardware_backend.npu.moe.activation import (
AllGatherActivationWrapper,
NPUGeluAndMul,
NPUSitu,
NP... | 351 | 12,006 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton_kernels.py | .py | """Triton kernels MoE runner backend skeleton."""
from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Optional
import torch
from sglang.srt.layers.moe.moe_runner.base import (
MoeQuantInfo,
MoeRunnerConfig,
MoeRunnerCore,
RunnerInput,
Runne... | 208 | 6,811 |
sglang | python/sglang/srt/layers/moe/moe_runner/flashinfer_trtllm.py | .py | from __future__ import annotations
import contextvars
from contextlib import contextmanager
from dataclasses import dataclass
from typing import TYPE_CHECKING, Generator, Optional, cast
import torch
from torch.nn import Module
from torch.nn.parameter import Parameter
from sglang.kernels.ops.moe.pack_topk_ids import ... | 1,386 | 53,711 |
sglang | python/sglang/srt/layers/moe/moe_runner/hpc_ops.py | .py | from __future__ import annotations
"""
MoE runner backend powered by HPC-Ops (https://github.com/Tencent/hpc-ops),
a production-grade operator library for LLM inference developed by the
Tencent Hunyuan AI Infra team.
The backend wraps the monolithic FP8 fused-MoE kernels ``fuse_moe_blockwise``
(128x128 block-quantize... | 209 | 7,936 |
sglang | python/sglang/srt/layers/moe/moe_runner/flashinfer_cutlass.py | .py | """FlashInfer CUTLASS MoE fused funcs.
This module owns the FlashInfer ``cutlass_fused_moe`` calls used by the
unquantized, ModelOpt FP8, ModelOpt NVFP4, and MXFP4 MoE paths.
Quantization methods prepare a small quant_info payload and route through
``MoeRunner``.
"""
from __future__ import annotations
from dataclass... | 399 | 14,142 |
sglang | python/sglang/srt/layers/moe/moe_runner/base.py | .py | from __future__ import annotations
from abc import ABC, abstractmethod
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Callable, Optional, Tuple, TypeGuard
import torch
from sglang.srt.layers.moe.utils import (
MoeA2ABackend,
MoeRunnerBackend,
RoutingMethodType,
)
if TYPE_CHECKI... | 305 | 9,734 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton.py | .py | from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, List, Optional
import torch
from sglang.srt.layers.moe.moe_runner.base import (
MoeQuantInfo,
MoeRunnerConfig,
MoeRunnerCore,
RunnerInput,
RunnerOutput,
register_fused_func,
registe... | 336 | 11,729 |
sglang | python/sglang/srt/layers/moe/moe_runner/humming.py | .py | from __future__ import annotations
import json
import logging
import math
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Optional
from weakref import WeakValueDictionary
import torch
from sglang.kernels.ops.moe.ep_moe_kernels import moe_permute, moe_unpermute
from sglang.kernels.ops.moe.moe... | 818 | 28,473 |
sglang | python/sglang/srt/layers/moe/moe_runner/aiter.py | .py | from __future__ import annotations
import functools
import inspect
from dataclasses import dataclass
from enum import Enum
from typing import TYPE_CHECKING, Any, Optional, Union
import torch
from sglang.srt.layers.moe.moe_runner.base import (
MoeQuantInfo,
MoeRunnerConfig,
MoeRunnerCore,
RunnerInput,... | 478 | 17,594 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton_utils/fused_moe.py | .py | # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# Adapted from https://github.com/vllm-project/vllm/blob/a6221a144af772fd1a68fe7e627935dc53e81738/vllm/model_executor/layers/fused_moe/fused_moe.py
"""Fused MoE kernel."""
from __future__ import annotations
imp... | 1,158 | 40,671 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton_utils/moe_align_block_size.py | .py | from __future__ import annotations
from typing import Tuple
import torch
import triton
from sglang.srt.environ import envs
from sglang.srt.utils import is_cuda, is_hip, is_musa, is_xpu
_SGLANG_EXPERIMENTAL_LORA_OPTI = envs.SGLANG_EXPERIMENTAL_LORA_OPTI.get()
_is_cuda = is_cuda()
_is_hip = is_hip()
_is_xpu = is_xpu... | 163 | 5,943 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton_utils/__init__.py | .py | from contextlib import contextmanager
from typing import Any, Dict, Optional
from sglang.srt.layers.moe.moe_runner.triton_utils.fused_moe import fused_experts
from sglang.srt.layers.moe.moe_runner.triton_utils.fused_moe_triton_config import (
get_config_file_name,
try_get_optimal_moe_config,
)
from sglang.srt.... | 37 | 837 |
sglang | python/sglang/srt/layers/moe/moe_runner/triton_utils/fused_moe_triton_config.py | .py | from __future__ import annotations
import functools
import json
import logging
import os
from typing import Any, Dict, List, Optional, Tuple
import torch
import triton
from sglang.srt.runtime_context import get_exec
from sglang.srt.utils import get_device_name, is_hip
logger = logging.getLogger(__name__)
_is_hip = ... | 363 | 12,850 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/deepep.py | .py | from __future__ import annotations
import logging
from contextlib import nullcontext
from dataclasses import dataclass
from typing import TYPE_CHECKING, List, NamedTuple, Optional, Tuple, Union
from sglang.srt.distributed.parallel_state import get_tp_group
from sglang.srt.environ import envs
from sglang.srt.eplb.expe... | 1,027 | 35,126 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/nixl.py | .py | from __future__ import annotations
import logging
from enum import Enum, auto
import torch
import torch.distributed as dist
from sglang.srt.distributed.utils import get_global_tcp_store
from sglang.srt.elastic_ep.elastic_ep import ElasticEPStateManager
from sglang.srt.environ import envs
from sglang.srt.eplb.expert_... | 554 | 18,356 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/ascend_tp.py | .py | from __future__ import annotations
from typing import NamedTuple, Optional
import torch
from sglang.srt.hardware_backend.npu.moe.finalize_routing import (
AllGatherFinalizeRoutingWrapper,
NPUFinalizeRouting,
)
from sglang.srt.hardware_backend.npu.moe.init_routing import (
MXFP8_QUANT_MODE,
NPUMoEInit... | 144 | 5,129 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/__init__.py | .py | from sglang.srt.layers.moe.token_dispatcher.ascend_tp import (
AscendTPCombineInput,
AscendTPDispatcher,
AscendTPDispatchOutput,
)
from sglang.srt.layers.moe.token_dispatcher.base import (
BaseDispatcher,
BaseDispatcherConfig,
CombineInput,
CombineInputChecker,
CombineInputFormat,
Di... | 94 | 2,470 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/mooncake.py | .py | from __future__ import annotations
import logging
from dataclasses import dataclass
from enum import Enum, auto
from typing import NamedTuple, Optional
import torch
import torch.distributed as dist
from sglang.srt.elastic_ep.elastic_ep import ElasticEPStateManager
from sglang.srt.eplb.expert_distribution import get_... | 410 | 12,713 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/flashinfer_utils.py | .py | import torch.distributed as dist
from sglang.srt.utils import is_flashinfer_available
if is_flashinfer_available():
from flashinfer.comm.mnnvl import CommBackend
else:
class CommBackend:
"""
Placeholder base class when flashinfer is not available
"""
pass
class TorchDistrib... | 48 | 1,265 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/moriep.py | .py | from __future__ import annotations
import logging
import os
from dataclasses import dataclass
from typing import TYPE_CHECKING, List, NamedTuple, Optional, Tuple
from sglang.srt.eplb.expert_distribution import (
_ExpertDistributionRecorderNoop,
get_global_expert_distribution_recorder,
)
from sglang.srt.layers... | 1,180 | 40,512 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/flashinfer.py | .py | from __future__ import annotations
import logging
from typing import NamedTuple, Optional
import torch
from sglang.kernel_api_logging import debug_kernel_api
from sglang.srt.environ import envs
from sglang.srt.layers.dp_attention import (
get_dp_global_num_tokens,
is_dp_attention_enabled,
)
from sglang.srt.l... | 317 | 13,565 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/pplx.py | .py | from __future__ import annotations
from enum import Enum, auto
from typing import NamedTuple, Optional, Tuple
import torch
import torch.distributed as dist
from sglang.srt.environ import envs
from sglang.srt.eplb.expert_distribution import get_global_expert_distribution_recorder
from sglang.srt.layers.dp_attention i... | 528 | 17,525 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/base.py | .py | from __future__ import annotations
import weakref
from abc import ABC, abstractmethod
from enum import Enum
from typing import (
TYPE_CHECKING,
Any,
Callable,
Optional,
OrderedDict,
Protocol,
Tuple,
TypeGuard,
Union,
runtime_checkable,
)
import torch
if TYPE_CHECKING:
from... | 392 | 12,449 |
sglang | python/sglang/srt/layers/moe/token_dispatcher/standard.py | .py | from __future__ import annotations
from typing import TYPE_CHECKING, NamedTuple, Optional, Tuple
import torch
from sglang.srt.distributed import (
get_tp_group,
)
from sglang.srt.distributed.device_communicators.pynccl_allocator import (
use_symmetric_memory,
)
from sglang.srt.layers.dp_attention import (
... | 261 | 10,457 |
sglang | python/sglang/srt/layers/utils/cp_utils.py | .py | from dataclasses import dataclass
from itertools import accumulate
from typing import Callable, List
import torch
import torch.nn.functional as F
from sglang.srt.distributed.device_communicators.pynccl_allocator import (
use_symmetric_memory,
)
from sglang.srt.layers.dp_attention import (
attn_cp_all_gather_i... | 676 | 24,929 |
sglang | python/sglang/srt/layers/utils/multi_platform.py | .py | """Deprecated compatibility shim for the former ``MultiPlatformOp``.
The multi-platform operator abstraction was unified into
:class:`sglang.kernels.fused_op.BaseFusedOp` (RFC #29630): one class now
covers kernel-backend selection (``forward_aot`` / ``forward_jit`` / ...),
platform dispatch (``forward_cuda`` / ``forwa... | 63 | 2,345 |
sglang | python/sglang/srt/layers/utils/common.py | .py | # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import logging
import re
import torch
from torch.nn.parameter import Parameter
logger = logging.getLogger(__name__)
def get_layer_id(weight_name):
# example weight name: model.layers.10.self_attn.qkv_proj.... | 128 | 4,584 |
sglang | python/sglang/srt/layers/deep_gemm_wrapper/entrypoint.py | .py | import logging
from contextlib import contextmanager
from typing import Any, Optional, Tuple
import torch
from sglang.srt.environ import envs
from sglang.srt.layers.deep_gemm_wrapper import compile_utils
from sglang.srt.layers.deep_gemm_wrapper.configurer import ( # noqa: F401
DEEPGEMM_BLACKWELL,
DEEPGEMM_NE... | 285 | 8,360 |
sglang | python/sglang/srt/layers/deep_gemm_wrapper/configurer.py | .py | import logging
from sglang.srt.environ import envs
from sglang.srt.utils import (
get_device_sm,
is_cuda,
is_musa,
is_sm100_supported,
)
logger = logging.getLogger(__name__)
_is_cuda = is_cuda()
_is_musa = is_musa()
def _compute_enable_deep_gemm():
sm_version = get_device_sm()
if (_is_cuda ... | 40 | 959 |
sglang | python/sglang/srt/layers/deep_gemm_wrapper/compile_utils.py | .py | import logging
import math
import os
import time
from contextlib import contextmanager, nullcontext
from enum import IntEnum, auto
from typing import Dict, List, Tuple
import torch
from tqdm import tqdm
from sglang.srt.distributed.device_communicators.pynccl_allocator import (
disable_symmetric_memory_context,
... | 511 | 19,700 |
sglang | python/sglang/srt/layers/attention/aiter_backend.py | .py | from __future__ import annotations
from sglang.srt.runtime_context import get_parallel, get_spec
"""
end to end attention solution with aiter kernels
"""
import logging
from dataclasses import dataclass
from enum import Enum, auto
from typing import TYPE_CHECKING, Optional
import torch
import triton
from sglang.ke... | 2,990 | 120,413 |
sglang | python/sglang/srt/layers/attention/torch_flex_backend.py | .py | from __future__ import annotations
from typing import TYPE_CHECKING
import torch
from torch.nn.attention.flex_attention import create_block_mask, flex_attention
from sglang.srt.layers.attention.base_attn_backend import AttentionBackend
from sglang.srt.layers.radix_attention import AttentionType
from sglang.srt.model... | 332 | 11,915 |
sglang | python/sglang/srt/layers/attention/aiter_utils.py | .py | """SHUFFLE 5D KV pool helpers for the AITER attention backend.
This module hosts the attention pathways that are specific to the
``SGLANG_AITER_KV_CACHE_LAYOUT=vectorized_5d`` (SHUFFLE 5D) physical layout.
They live here rather than inline in
:mod:`sglang.srt.layers.attention.aiter_backend` so the main backend
file ke... | 310 | 11,970 |
sglang | python/sglang/srt/layers/attention/dsa_backend.py | .py | from __future__ import annotations
import logging
from dataclasses import dataclass
from typing import (
TYPE_CHECKING,
Dict,
List,
Literal,
Optional,
Tuple,
TypeAlias,
)
import torch
from sglang.srt.configs.model_config import get_dsa_index_topk, is_deepseek_dsa
from sglang.srt.runtime_c... | 3,640 | 153,237 |
sglang | python/sglang/srt/layers/attention/intel_amx_backend.py | .py | from __future__ import annotations
from typing import TYPE_CHECKING
import torch
from sglang.srt.layers.attention.base_attn_backend import AttentionBackend
from sglang.srt.mem_cache.memory_pool import KVWriteLoc
from sglang.srt.mem_cache.swa_memory_pool import SWAKVPool
from sglang.srt.model_executor.forward_batch_i... | 362 | 13,896 |
sglang | python/sglang/srt/layers/attention/cutlass_mla_backend.py | .py | from __future__ import annotations
from sglang.srt.runtime_context import get_parallel
"""
Support attention backend for Cutlass MLA.
"""
from dataclasses import dataclass
from typing import TYPE_CHECKING, Optional, Union
import torch
import triton
from sglang.kernels.ops.attention.utils import (
create_flash... | 251 | 8,708 |
sglang | python/sglang/srt/layers/attention/triton_backend.py | .py | from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, List, Optional
import torch
import triton
from sglang.kernels.ops.attention.metadata import get_num_kv_splits_triton
from sglang.kernels.ops.kvcache.kv_indices import (
create_flashinfer_kv_indices_triton,
)
fr... | 2,170 | 88,099 |
sglang | python/sglang/srt/layers/attention/index_topk_share.py | .py | from __future__ import annotations
from contextlib import contextmanager
from typing import TYPE_CHECKING, Iterator, Optional
if TYPE_CHECKING:
import torch
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
class IndexTopKShareState:
def __init__(
self,
forward_batch... | 87 | 2,810 |
sglang | python/sglang/srt/layers/attention/deepseek_v4_backend.py | .py | from __future__ import annotations
import enum
import functools
import logging
from dataclasses import dataclass, field
from typing import (
TYPE_CHECKING,
Dict,
List,
Literal,
Optional,
Tuple,
TypeVar,
Union,
)
import torch
import torch.nn.functional as F
from sglang.kernels.ops.atte... | 2,212 | 90,625 |
sglang | python/sglang/srt/layers/attention/merge_state.py | .py | from typing import Optional, Tuple
import torch
from sgl_kernel import merge_state_v2
from sglang.kernels.ops.attention.merge_state import merge_state_triton
from sglang.srt.utils import is_cuda
_is_cuda = is_cuda()
# Automatically fallback to the Triton kernel in some cases
# (e.g., for AMD GPUs, when the head di... | 47 | 1,413 |
sglang | python/sglang/srt/layers/attention/xpu_backend.py | .py | from __future__ import annotations
from typing import TYPE_CHECKING, Optional
import torch
from sglang.srt.configs.model_config import AttentionArch
from sglang.srt.layers.attention.base_attn_backend import AttentionBackend
from sglang.srt.layers.attention.flashattention_backend import (
FlashAttentionMetadata,
... | 1,312 | 60,721 |
sglang | python/sglang/srt/layers/attention/trtllm_mha_backend.py | .py | from __future__ import annotations
"""
Support attention backend for TRTLLM MHA kernels from flashinfer.
The kernel supports sm100 only, with sliding window and attention sink features.
"""
import logging
from dataclasses import dataclass
from typing import TYPE_CHECKING, Optional
import torch
from sglang.kernels.o... | 1,476 | 62,369 |
sglang | python/sglang/srt/layers/attention/unified_mem_hooks.py | .py | # Copyright 2023-2026 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writin... | 71 | 2,666 |
sglang | python/sglang/srt/layers/attention/hybrid_linear_attn_backend.py | .py | from __future__ import annotations
import logging
from typing import TYPE_CHECKING, Optional, Union
import torch
from sglang.kernels.ops.mamba.causal_conv1d_triton import PAD_SLOT_ID
from sglang.kernels.ops.mamba.mamba_state_indices_triton import (
fused_replay_state_indices,
)
from sglang.kernels.ops.mamba.mamb... | 1,281 | 58,018 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.