MCPcopy Create free account

hub / github.com/UbiquitousLearning/mllm / functions

Functions5,971 in github.com/UbiquitousLearning/mllm

⨍ Functions 5,971◇ Types & classes 1,855↳ Endpoints 4
↓ 3 callersMethodgetDevice
mllm/core/BaseOp.cpp:20
↓ 3 callersMethodgetExtraTensorViewInTensor
mllm/core/Tensor.cpp:113
↓ 3 callersMethodgetGraphOutputTensorWrappers
mllm/backends/qnn/QNNModel.hpp:75
↓ 3 callersMethodgetHiddenStateCache
algorithms/lazy_vlm/HKVCache.cpp:162
↓ 3 callersFunctiongetInfoHelper
mllm/backends/opencl/vendors/OpenCL-Headers/CL/opencl.hpp:1116
↓ 3 callersMethodgetLayerNums
mllm/nn/lmcache/StaticCache.hpp:51
↓ 3 callersMethodgetNextToken
mllm/compile/symbolic_expr/Parser.cpp:56
↓ 3 callersFunctiongetPageAddr
mllm/engine/prefix_cache/TLB.cpp:12
↓ 3 callersMethodgetQnnGraphName
mllm/backends/qnn/QNNModel.hpp:71
↓ 3 callersFunctionget_scale_min_k4
mllm/backends/cpu/kernels/common/ggml/quantize/quantize_q4.cpp:173
↓ 3 callersMethodhasAttachedView
mllm/core/Tensor.hpp:721
↓ 3 callersMethodhash
mllm/core/Tensor.cpp:538
↓ 3 callersMethodinitAttentionMask
mllm/backends/qnn/aot_rt/KVCacheManager.cpp:68
↓ 3 callersMethodinitHiddenStateCache
algorithms/lazy_vlm/HKVCache.cpp:152
↓ 3 callersMethodinitialize
mllm/backends/qnn/QNNModel.cpp:19
↓ 3 callersMethodinsertOpAtLast
mllm/compile/ir/Node.cpp:385
↓ 3 callersFunctioninterpolateBySize
mllm/nn/Functional.cpp:165
↓ 3 callersMethodintrusive_ptr_add_ref
mllm/utils/IntrusivePtr.hpp:23
↓ 3 callersFunctionirange
mllm-kernel/include/mllm_kernel/utils.hpp:155
↓ 3 callersMethodisPositional
mllm/utils/Argparse.hpp:103
↓ 3 callersMethodisStatic
mllm/compile/symbolic_expr/AST.hpp:205
↓ 3 callersFunctionis_power_of_2
mllm/backends/cpu/ops/STFTOp.cpp:34
↓ 3 callersMethodlangle
mllm/compile/ir/IRPrinter.hpp:29
↓ 3 callersMethodlbrace
mllm/compile/ir/IRPrinter.hpp:18
↓ 3 callersFunctionlog_ps_f32
FIXME: The code below is from LLM. Have precision errors.
mllm/backends/cpu/kernels/arm/primitives.hpp:388
↓ 3 callersMethodlookup
mllm/compile/symbolic_expr/Parser.cpp:35
↓ 3 callersMethodmatch
tools/mllm-quantizer/schema/kai.cpp:12
↓ 3 callersMethodmatch
(self, quantize_config, tensor_dict: Dict, **kwargs)
pymllm/quantize/quantize_pass.py:26
↓ 3 callersFunctionmllm_blas_batch_matmul_fp32
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:446
↓ 3 callersFunctionmllm_blas_matmul_fp32
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:414
↓ 3 callersFunctionmllm_vld1q_s8_x4
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:506
↓ 3 callersFunctionmllm_vld1q_u8_x2
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:465
↓ 3 callersFunctionmul_sum_i8_pairs_float
multiply int8_t, add results pairwise twice and return as float vector
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:274
↓ 3 callersMethodnewline
mllm/compile/ir/IRPrinter.hpp:14
↓ 3 callersFunctionpacked_weights_size_per_n
Size in bytes of 1 packed weights column
mllm/backends/cpu/kernels/arm/quantize/bitspack/bitspack.hpp:238
↓ 3 callersMethodparse
mllm/utils/Argparse.hpp:70
↓ 3 callersMethodperfSummary
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:1561
↓ 3 callersMethodquant_pack_rhs_offline
mllm/backends/cpu/kernels/arm/linear/kai.cpp:62
↓ 3 callersMethodquant_pack_rhs_size
mllm/backends/cpu/kernels/arm/linear/kai.cpp:54
↓ 3 callersMethodrangle
mllm/compile/ir/IRPrinter.hpp:30
↓ 3 callersMethodrbrace
mllm/compile/ir/IRPrinter.hpp:23
↓ 3 callersMethodread
===----------------------------------------------------------------------===// CPU ModelFileV1 ParameterFile ===--------------------------------------
mllm/core/ParameterFile.cpp:68
↓ 3 callersMethodremove
mllm/engine/prefix_cache/TLB.cpp:24
↓ 3 callersMethodresetCache
Reset KV cache
mllm/models/smollm3_3B/modeling_smollm3.hpp:274
↓ 3 callersMethodresetRegion
mllm/compile/ir/Node.cpp:242
↓ 3 callersFunctionrmsnorm_fp32
Should support [B, S, H * D] and [B, S, H, D]
mllm/backends/cpu/kernels/arm/rmsnorm.cpp:13
↓ 3 callersMethodrun
(self, quantize_config, tensor_dict: Dict, **kwargs)
pymllm/quantize/quantize_pass.py:30
↓ 3 callersFunctionscaledDotProductAttention
mllm/nn/Functional.cpp:151
↓ 3 callersFunctionsendRequest
mllm/c_api/Runtime.cpp:141
↓ 3 callersMethodsetLayerIndex
mllm/nn/layers/KVCache.cpp:20
↓ 3 callersMethodsize
mllm/core/TensorViewImpl.cpp:27
↓ 3 callersMethodstreamGenerate
mllm/models/ARGeneration.cpp:200
↓ 3 callersMethodstream_quantize
( self, quant_cfg, tensor_dict: Dict, writer: ModelFileV2, **kwargs )
pymllm/quantize/solver.py:77
↓ 3 callersMethodstream_quantize_params_size
( self, quant_cfg, tensor_dict: Dict, **kwargs )
pymllm/quantize/solver.py:25
↓ 3 callersMethodstreaming_write
(self, tensor_name, tensor_obj)
pymllm/convertor/model_file_v2.py:139
↓ 3 callersMethodswap
mllm/utils/AnyValue.hpp:114
↓ 3 callersFunctiontraceContinue
mllm/compile/ir/Trace.cpp:109
↓ 3 callersFunctiontraceModule
mllm/compile/ir/Trace.hpp:42
↓ 3 callersFunctiontraceYield
mllm/compile/ir/Trace.cpp:103
↓ 3 callersMethodupdateAttentionMask
mllm/backends/qnn/aot_rt/KVCacheManager.cpp:146
↓ 3 callersFunctionvexpq_fast_f16
mllm/backends/cpu/kernels/arm/math.hpp:393
↓ 3 callersFunctionvpaddq_s32
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:382
↓ 2 callersMethodInsert
(sessionID string)
mllm-cli/mllm/c.go:102
↓ 2 callersFunctionQDQ_KV
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot_sha.hpp:82
↓ 2 callersFunctionQDQ_KV
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot.hpp:77
↓ 2 callersFunctionQDQ_KV
examples/llama_qnn_aot/modeling_llama_qnn_aot_sha.hpp:82
↓ 2 callersFunctionQDQ_KV
examples/llama_qnn_aot/modeling_llama_qnn_aot.hpp:77
↓ 2 callersFunctionQDQ_KV
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot_sha.hpp:82
↓ 2 callersFunctionQDQ_KV
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot.hpp:77
↓ 2 callersFunctionQDQ_ROPE
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot_sha.hpp:107
↓ 2 callersFunctionQDQ_ROPE
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot.hpp:102
↓ 2 callersFunctionQDQ_ROPE
examples/llama_qnn_aot/modeling_llama_qnn_aot_sha.hpp:107
↓ 2 callersFunctionQDQ_ROPE
examples/llama_qnn_aot/modeling_llama_qnn_aot.hpp:102
↓ 2 callersFunctionQDQ_ROPE
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot_sha.hpp:107
↓ 2 callersFunctionQDQ_ROPE
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot.hpp:102
↓ 2 callersMethodRegisterSession
(id string, session *mllm.Session)
mllm-cli/pkg/mllm/service.go:23
↓ 2 callersMethod__forceSetDevice
used for set default device when creating customized ops
mllm/nn/AbstractNnNode.hpp:58
↓ 2 callersMethod__forceSetOpType
mllm/nn/Layer.hpp:38
↓ 2 callersFunction__from_dlpack
mllm/ffi/Object.cc:16
↓ 2 callersMethod__init__
( self, dtype=torch.uint8, qscheme=torch.per_tensor_affine, reduce_range=False
pymllm/backends/qualcomm/transformers/core/observer.py:152
↓ 2 callersMethod__init__
(self, in_features, out_features, bias=True)
pymllm/backends/qualcomm/transformers/core/qlinear.py:15
↓ 2 callersMethod__init__
(self)
pymllm/nn/_layers.py:70
↓ 2 callersMethod__main
mllm/nn/Layer.cpp:53
↓ 2 callersFunction__mllm_blas_batch_matmul_fp32_gemv
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:212
↓ 2 callersFunction__mllm_blas_matmul_fp32_gemv
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:200
↓ 2 callersFunction__mllm_blas_matmul_fp32_gemv_nt_t_decode_small_d_qk_baseline
Optimized for decoding. Q: [1, D] K: [S, D] D is small in mllm's case(small language model). D=64, 96, 128 ...
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:18
↓ 2 callersFunction__mllm_blas_sgemm_nt_nt
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:368
↓ 2 callersFunction__mllm_blas_sgemm_nt_t
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_sgemm.cpp:391
↓ 2 callersMethod__setDataContainer
mllm/backends/qnn/QNNUtils.hpp:211
↓ 2 callersMethod__trace
mllm/nn/Module.cpp:144
↓ 2 callersMethod_param_name_view
mllm/core/schema/ModelFileV2.hpp:57
↓ 2 callersFunction_recursiveSolveNormalImpl
mllm/backends/qnn/aot/passes/PTQPass.cpp:156
↓ 2 callersMethod_repr_helper
(self, indent_level: int = 0)
pymllm/nn/_module.py:75
↓ 2 callersFunction_resolve_kernel_path
Resolve the path to the mllm_kernel package directory.
mllm-kernel/mllm_kernel/jit_utils/compile.py:54
↓ 2 callersFunction_run_command
(args: list[str])
mllm-kernel/mllm_kernel/__main__.py:117
↓ 2 callersFunction_torch_tensor_bytes
(tensor: "torch.Tensor")
pymllm/convertor/model_file_v2.py:27
↓ 2 callersFunction_vec_clip_inplace
mllm/backends/cpu/kernels/arm/quantize/bitspack/bitspack.cpp:93
↓ 2 callersMethodaddTuningParameter
mllm/experiments/auto_tune/TuningSpace.cpp:49
↓ 2 callersFunctionadd_constant_runtime_simd_impl
mllm-kernel/mllm_kernel/cpu/include/mllm_kernel/cpu/add_constant.hpp:13
↓ 2 callersMethodalloc
mllm/backends/ascend/AscendAllocator.cpp:23
← previousnext →501–600 of 5,971, ranked by callers