MCPcopy Create free account

hub / github.com/UbiquitousLearning/mllm / functions

Functions5,971 in github.com/UbiquitousLearning/mllm

↓ 2 callersFunctionreduce_sum_int8
mllm/backends/cpu/kernels/arm/reduce.cpp:66
↓ 2 callersMethodregisterSession
mllm/engine/service/Service.cpp:82
↓ 2 callersMethodrelease
mllm/backends/ascend/AscendCommon.cpp:43
↓ 2 callersMethodremoveAttr
mllm/compile/ir/Node.cpp:57
↓ 2 callersMethodremoveOpWithoutEdgeCut
mllm/compile/ir/Node.cpp:318
↓ 2 callersFunctionrender_flags
(flags: list[str])
mllm-kernel/mllm_kernel/__main__.py:80
↓ 2 callersMethodreorderKVCache
algorithms/lazy_vlm/HKVCache.cpp:194
↓ 2 callersFunctionrepeat_kv
This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he
pymllm/backends/qualcomm/transformers/qwen3/modeling_qwen3.py:141
↓ 2 callersFunctionrepeat_kv
This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he
pymllm/backends/qualcomm/transformers/qwen2/modeling_qwen2.py:126
↓ 2 callersFunctionreplicate_index
Helper: replicate index (clamp)
mllm/backends/cpu/ops/PadOp.cpp:40
↓ 2 callersFunctionresample
mllm/preprocessor/audio/Audio.cpp:341
↓ 2 callersMethodrescale
mllm/backends/cpu/kernels/common/fa2/fwd_bshd.hpp:179
↓ 2 callersMethodrewrite
mllm/backends/qnn/aot/passes/LLMQuantRecipePass.cpp:264
↓ 2 callersFunctionrotateHalf
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot_sha.hpp:131
↓ 2 callersFunctionrotateHalf
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot.hpp:126
↓ 2 callersFunctionrotateHalf
examples/llama_qnn_aot/modeling_llama_qnn_aot_sha.hpp:131
↓ 2 callersFunctionrotateHalf
examples/llama_qnn_aot/modeling_llama_qnn_aot.hpp:126
↓ 2 callersFunctionrotateHalf
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot_sha.hpp:131
↓ 2 callersFunctionrotateHalf
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot.hpp:126
↓ 2 callersMethodrun
tools/mllm-llm-benchmark/models/Qwen3_W4A32_KAI.hpp:100
↓ 2 callersMethodrun
mllm/compile/PassManager.cpp:26
↓ 2 callersMethodsampleGreedy
mllm/backends/qnn/aot_rt/QnnAOTModule.cpp:17
↓ 2 callersMethodsampleTemperature
mllm/models/ARGeneration.cpp:349
↓ 2 callersMethodsampleTopK
mllm/models/ARGeneration.cpp:369
↓ 2 callersMethodsampleTopP
mllm/models/ARGeneration.cpp:407
↓ 2 callersMethodselect_high_score_visual_token_decode
algorithms/lazy_vlm/models/qwen2_5vl/lazy_vlm_cfg.hpp:127
↓ 2 callersMethodselect_high_score_visual_token_prefill
algorithms/lazy_vlm/models/qwen2_5vl/lazy_vlm_cfg.hpp:87
↓ 2 callersMethodsetCtx
mllm/compile/passes/Pass.cpp:10
↓ 2 callersMethodsetInplace
mllm/core/BaseOp.hpp:66
↓ 2 callersMethodsetInstancedOp
mllm/nn/Layer.cpp:47
↓ 2 callersMethodsetMappedFile
mllm/core/ParameterFile.cpp:24
↓ 2 callersMethodsetProgramIntrinsicId
mllm/compile/ir/program/Op.hpp:27
↓ 2 callersMethodsetQNNPointer
mllm/backends/qnn/QNNAllocator.hpp:64
↓ 2 callersMethodsetSVMPointers
! * Specify a vector of SVM pointers that the kernel may access in * addition to its arguments. */
mllm/backends/opencl/vendors/OpenCL-Headers/CL/opencl.hpp:6435
↓ 2 callersMethodsetStaticCache
mllm/nn/layers/KVCache.cpp:34
↓ 2 callersMethodsetValueName
mllm/compile/ir/Node.cpp:228
↓ 2 callersMethodsetup
mllm/backends/opencl/ops/X2XOp.cpp:15
↓ 2 callersFunctionshow_env
()
mllm-kernel/mllm_kernel/__main__.py:104
↓ 2 callersMethodshutdown
mllm/engine/service/Service.cpp:35
↓ 2 callersMethodshutdown
mllm/backends/qnn/QNNBackend.cpp:183
↓ 2 callersFunctionsigmoid_fp32
mllm/backends/cpu/kernels/arm/sigmoid.cpp:13
↓ 2 callersFunctionsilu_fp32
mllm/backends/cpu/kernels/arm/silu.cpp:13
↓ 2 callersFunctionsoftmax_v1_fp16
mllm/backends/cpu/kernels/arm/softmax.cpp:156
↓ 2 callersFunctionsolveLinearWeight
mllm/backends/qnn/aot/passes/PTQPass.cpp:33
↓ 2 callersMethodstart
===----------------------------------------------------------------------===// Service ===------------------------------------------------------------
mllm/engine/service/Service.cpp:93
↓ 2 callersFunctionstartService
mllm/c_api/Runtime.cpp:66
↓ 2 callersMethodstop
mllm/engine/service/Service.cpp:99
↓ 2 callersFunctionstopService
mllm/c_api/Runtime.cpp:71
↓ 2 callersMethodstore
mllm/backends/cpu/kernels/arm/primitives.hpp:555
↓ 2 callersMethodtoFloat
mllm/preprocessor/audio/Audio.cpp:105
↓ 2 callersFunctiontoMono
Convert stereo audio data to mono by averaging channels
mllm/preprocessor/audio/Audio.cpp:425
↓ 2 callersMethodto_pod
(self)
pymllm/ffi/__init__.py:47
↓ 2 callersMethodtokenize
mllm/models/minicpm_o2_6/tokenization_chattts.cpp:61
↓ 2 callersMethodtokenize
mllm/preprocessor/tokenizers/WordPiece.cpp:151
↓ 2 callersFunctiontrace
mllm/compile/ir/Trace.hpp:15
↓ 2 callersFunctiontranspose
mllm/preprocessor/audio/Audio.cpp:146
↓ 2 callersFunctionunicode_cpt_from_utf8
mllm/preprocessor/tokenizers/llama_cpp_unicode/unicode.cpp:76
↓ 2 callersFunctionunicode_regex_split
mllm/preprocessor/tokenizers/llama_cpp_unicode/unicode.cpp:657
↓ 2 callersFunctionunicode_regex_split_stl
use std::wregex to split the text
mllm/preprocessor/tokenizers/llama_cpp_unicode/unicode.cpp:498
↓ 2 callersFunctionunicode_tolower
mllm/preprocessor/tokenizers/llama_cpp_unicode/unicode.cpp:649
↓ 2 callersFunctionunicode_utf8_to_byte
mllm/preprocessor/tokenizers/llama_cpp_unicode/unicode.cpp:644
↓ 2 callersMethodupdateCache
mllm/backends/qnn/aot_rt/KVCacheManager.cpp:242
↓ 2 callersFunctionvec_clamp
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_i8gemm_bitspack_1x8x16_fp32_neondot.hpp:18
↓ 2 callersFunctionvec_pack_128_uint1_values
This function is a vectorized version of pack_8_uint1_values To understand it, please see `pack_8_uint1_values` first. Input is 128 bytes Output is 1
mllm/backends/cpu/kernels/arm/quantize/bitspack/u1.hpp:82
↓ 2 callersFunctionvec_pack_128_uint3_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u3.hpp:165
↓ 2 callersFunctionvec_pack_128_uint5_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u5.hpp:113
↓ 2 callersFunctionvec_pack_128_uint7_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u7.hpp:159
↓ 2 callersFunctionvec_pack_32_uint2_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u2.hpp:39
↓ 2 callersFunctionvec_pack_32_uint6_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u6.hpp:53
↓ 2 callersFunctionvec_pack_64_uint1_values
This function is a vectorized version of pack_8_uint1_values To understand it, please see pack_8_uint1_values first. Input is 64 bytes Output is 64 b
mllm/backends/cpu/kernels/arm/quantize/bitspack/u1.hpp:43
↓ 2 callersFunctionvec_pack_64_uint3_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u3.hpp:70
↓ 2 callersFunctionvec_pack_64_uint5_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u5.hpp:62
↓ 2 callersFunctionvec_pack_64_uint7_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u7.hpp:87
↓ 2 callersFunctionvec_unpack_128_uint1_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u1.hpp:101
↓ 2 callersFunctionvec_unpack_128_uint3_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u3.hpp:205
↓ 2 callersFunctionvec_unpack_128_uint5_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u5.hpp:147
↓ 2 callersFunctionvec_unpack_128_uint7_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u7.hpp:189
↓ 2 callersFunctionvec_unpack_32_uint2_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u2.hpp:56
↓ 2 callersFunctionvec_unpack_32_uint6_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u6.hpp:91
↓ 2 callersFunctionvec_unpack_64_uint1_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u1.hpp:60
↓ 2 callersFunctionvec_unpack_64_uint3_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u3.hpp:110
↓ 2 callersFunctionvec_unpack_64_uint5_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u5.hpp:89
↓ 2 callersFunctionvec_unpack_64_uint7_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/u7.hpp:117
↓ 2 callersFunctionvexpq_fast_f32
mllm/backends/cpu/kernels/x86/math.hpp:19
↓ 2 callersFunctionvexpq_hp_f32
mllm/backends/cpu/kernels/arm/math.hpp:29
↓ 2 callersFunctionvisitCallGraph
mllm/backends/qnn/passes/QNNOpNamingPass.cpp:61
↓ 2 callersFunctionvisitCallGraph
mllm/backends/qnn/passes/QNNGraphIOTensorPass.cpp:48
↓ 2 callersFunctionvisitCallGraph
mllm/backends/qnn/aot/passes/OpNamingPass.cpp:49
↓ 2 callersFunctionvisitCallGraph
mllm/backends/qnn/aot/passes/LPBQCanonicalizePass.cpp:23
↓ 2 callersFunctionvpaddq_s16
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:376
↓ 2 callersFunctionwait
mllm/mllm.hpp:106
↓ 2 callersMethodwalk
mllm/compile/ir/Node.hpp:516
↓ 2 callersFunctionwaveClip
mllm/preprocessor/audio/Audio.cpp:120
↓ 2 callersMethodwhat
! \brief Get error string associated with exception * * \return A memory pointer to the error message string. */
mllm/backends/opencl/vendors/OpenCL-Headers/CL/opencl.hpp:766
↓ 2 callersFunctionwrite_be64
mllm/engine/prefix_cache/HaHaHash.cpp:15
↓ 1 callersMethodContext
mllm/engine/Context.cpp:19
↓ 1 callersMethodGetSession
(sessionID string)
mllm-cli/pkg/mllm/service.go:29
↓ 1 callersMethodHtpBurst
()
pymllm/ffi/__init__.py:528
↓ 1 callersMethodHtpUnsignedPd
()
pymllm/ffi/__init__.py:575
↓ 1 callersFunctionImage2DGL
\brief Default constructor - initializes to nullptr.
mllm/backends/opencl/vendors/OpenCL-Headers/CL/opencl.hpp:5467
← previousnext →801–900 of 5,971, ranked by callers