MCPcopy Create free account

hub / github.com/UbiquitousLearning/mllm / functions

Functions5,971 in github.com/UbiquitousLearning/mllm

↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/qwen3/modeling_qwen3_service.hpp:34
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/qwen3/modeling_qwen3_qnn_aot.hpp:22
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/qwen3/modeling_qwen3.hpp:22
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/llama/modeling_llama.hpp:23
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/llama/modeling_llama_opencl.hpp:24
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/smollm3_3B/modeling_smollm3.hpp:26
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/deepseek_ocr/modeling_deepseek_ocr.hpp:32
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/qwen_npu/modeling_qwen_npu_cpu.hpp:28
↓ 1 callersFunctionmakeRotaryPosEmbedding
mllm/models/qwen_npu/modeling_qwen_npu.hpp:26
↓ 1 callersFunctionmakeVisualRoPEInvFreq
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:89
↓ 1 callersFunctionmakeVisualRoPEInvFreq
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:91
↓ 1 callersFunctionmakeVisualRoPEInvFreq
mllm/models/minicpm_o2_6/modeling_qwen2vl_for_minicpmo.hpp:17
↓ 1 callersFunctionmakeVisualRoPEInvFreq
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:108
↓ 1 callersFunctionmakeVisualRoPEInvFreq
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:108
↓ 1 callersFunctionmakeVisualRoPEInvFreq
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:37
↓ 1 callersFunctionmakeVisualRotaryPosEmb
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:203
↓ 1 callersFunctionmakeVisualRotaryPosEmb
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:205
↓ 1 callersFunctionmakeVisualRotaryPosEmb
mllm/models/minicpm_o2_6/modeling_qwen2vl_for_minicpmo.hpp:131
↓ 1 callersFunctionmakeVisualRotaryPosEmb
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:222
↓ 1 callersFunctionmakeVisualRotaryPosEmb
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:222
↓ 1 callersFunctionmakeVisualRotaryPosEmb
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:151
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:167
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:169
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
mllm/models/minicpm_o2_6/modeling_qwen2vl_for_minicpmo.hpp:95
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:186
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:186
↓ 1 callersFunctionmakeVisualRotaryPosEmbFull
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:115
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:101
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:103
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
mllm/models/minicpm_o2_6/modeling_qwen2vl_for_minicpmo.hpp:29
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:120
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:120
↓ 1 callersFunctionmakeVisualRotaryPosEmbIds
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:49
↓ 1 callersFunctionmakeVisualRotarySinCos
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:181
↓ 1 callersFunctionmakeVisualRotarySinCos
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:183
↓ 1 callersFunctionmakeVisualRotarySinCos
mllm/models/minicpm_o2_6/modeling_qwen2vl_for_minicpmo.hpp:109
↓ 1 callersFunctionmakeVisualRotarySinCos
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:200
↓ 1 callersFunctionmakeVisualRotarySinCos
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:200
↓ 1 callersFunctionmakeVisualRotarySinCos
algorithms/lazy_vlm/models/qwen2vl/modeling_qwen2vl.hpp:129
↓ 1 callersFunctionmakeWindowIndex
mllm/models/qwen2_5vl/modeling_qwen2_5vl_traceable.hpp:18
↓ 1 callersFunctionmakeWindowIndex
mllm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:20
↓ 1 callersFunctionmakeWindowIndex
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl.hpp:37
↓ 1 callersFunctionmakeWindowIndex
algorithms/lazy_vlm/models/qwen2_5vl/modeling_qwen2_5vl_fast.hpp:37
↓ 1 callersFunctionmake_block_q4_0x4
interleave 4 block_q4_0s in blocks of blck_size_interleave returns an interleaved block_q4_0x4 in the interleaved block_q4_0x4, place deltas for 4 blo
mllm/backends/cpu/kernels/arm/linear/ggml/gemm_aarch64.cpp:60
↓ 1 callersFunctionmake_block_q4_0x8
interleave 8 block_q4_0s in blocks of blck_size_interleave returns an interleaved block_q4_0x8 in the interleaved block_q4_0x8, place deltas for 8 blo
mllm/backends/cpu/kernels/arm/linear/ggml/gemm_aarch64.cpp:81
↓ 1 callersFunctionmake_cpp_args
Convert Python arguments to C++ template argument strings. Example: make_cpp_args(16, torch.float16) -> ["16", "fp16_t"]
mllm-kernel/mllm_kernel/jit_utils/compile.py:146
↓ 1 callersFunctionmake_q3_quants
mllm/backends/cpu/kernels/common/ggml/quantize/quantize_q3.cpp:34
↓ 1 callersFunctionmake_qkx2_quants
mllm/backends/cpu/kernels/common/ggml/quantize/quantize_q2.cpp:32
↓ 1 callersFunctionmake_qkx2_quants
mllm/backends/cpu/kernels/common/ggml/quantize/quantize_q4.cpp:97
↓ 1 callersFunctionmake_qp_quants
mllm/backends/cpu/kernels/common/ggml/quantize/quantize_q2.cpp:228
↓ 1 callersFunctionmake_qx_quants
mllm/backends/cpu/kernels/common/ggml/quantize/quantize.hpp:269
↓ 1 callersMethodmalloc
mllm/engine/prefix_cache/ZenFS.cpp:563
↓ 1 callersFunctionmap_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/bitspack.hpp:261
↓ 1 callersFunctionmaskedScatter
mllm/nn/Functional.cpp:185
↓ 1 callersMethodmaxWorkGroupSize
mllm/backends/opencl/runtime/OpenCLRuntime.cpp:92
↓ 1 callersMethodmean
(self, dim: int = -1, keep_dim: bool = False)
pymllm/ffi/__init__.py:202
↓ 1 callersFunctionmel_to_hertz_slaney
Convert frequency from mel scale to hertz (Slaney implementation)
mllm/models/minicpm_o2_6/audio_preprocessor_minicpmo.hpp:31
↓ 1 callersFunctionminiCPMORegex
mllm/models/minicpm_o2_6/tokenization_minicpmo.hpp:141
↓ 1 callersFunctionminiCPMOTokenizerMatchPattern
same with qwen2 参考: https://github.com/QwenLM/Qwen2-VL/blob/main/qwen2_vl/tokenization_qwen2_vl.py (?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L
mllm/models/minicpm_o2_6/tokenization_minicpmo.hpp:24
↓ 1 callersFunctionmllm_add_fp32
For arithmetic functions
mllm/backends/cpu/kernels/common/ggml/matmul.cpp:13
↓ 1 callersFunctionmllm_cpu_get_sve_cnt
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:335
↓ 1 callersFunctionmllm_gelu_f32
mllm/backends/cpu/kernels/common/ggml/quantize/quantize.hpp:165
↓ 1 callersFunctionmllm_gelu_quick_f32
mllm/backends/cpu/kernels/common/ggml/quantize/quantize.hpp:169
↓ 1 callersFunctionmllm_silu_f32
Sigmoid Linear Unit (SiLU) function
mllm/backends/cpu/kernels/common/ggml/quantize/quantize.hpp:172
↓ 1 callersFunctionmllm_vld1q_s16_x2
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:452
↓ 1 callersFunctionmllm_vld1q_s8_x2
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:493
↓ 1 callersFunctionmul
mllm/backends/cpu/kernels/common/llamafile/llamafile_sgemm.cpp:81
↓ 1 callersMethodmul_
mllm/core/Tensor.cpp:165
↓ 1 callersFunctionmul_sum_us8_pairs_float
mllm/backends/cpu/kernels/common/ggml/vec_dot.hpp:256
↓ 1 callersFunctionmyFunction
examples/tracy_example/main.cpp:5
↓ 1 callersFunctionnaive_conv2d
tests/cpu/Conv2DKernelTest.hpp:13
↓ 1 callersFunctionopencl_codegen
()
mllm/backends/opencl/kernels/gen_opencl_kernel_cpp_wrap.py:20
↓ 1 callersFunctionpack_values
mllm/backends/cpu/kernels/arm/quantize/bitspack/ux_packing.hpp:23
↓ 1 callersFunctionpacked_activations_size
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_i8gemm.hpp:26
↓ 1 callersFunctionpacked_weights_size
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_i8gemm.hpp:49
↓ 1 callersFunctionpacked_weights_with_lut_size
mllm/backends/cpu/kernels/arm/mllm_blas/mllm_blas_i8gemm.hpp:91
↓ 1 callersFunctionpad
mllm/nn/Functional.cpp:160
↓ 1 callersFunctionpadSignal
mllm/backends/cpu/ops/STFTOp.cpp:279
↓ 1 callersFunctionparseChipset
Helper function to parse chipset from string
mllm/backends/qnn/aot/QnnTargetMachineParser.cpp:26
↓ 1 callersFunctionparseHTPArch
Helper function to parse HTP architecture from string
mllm/backends/qnn/aot/QnnTargetMachineParser.cpp:14
↓ 1 callersFunctionparsePerformance
Helper function to parse performance mode from string
mllm/backends/qnn/aot/QnnTargetMachineParser.cpp:50
↓ 1 callersFunctionparseQcomTargetMachineFromJSON
mllm/backends/qnn/aot/QnnTargetMachineParser.cpp:70
↓ 1 callersFunctionparseSecurityPDSession
Helper function to parse security PD session from string
mllm/backends/qnn/aot/QnnTargetMachineParser.cpp:64
↓ 1 callersMethodperform
tools/mllm-quantizer/schema/kai.cpp:25
↓ 1 callersMethodphysicalAddr
mllm/engine/prefix_cache/Allocator.cpp:16
↓ 1 callersMethodphysicalAddrKey
mllm/nn/lmcache/PrefixCache.cpp:32
↓ 1 callersMethodphysicalAddrValue
mllm/nn/lmcache/PrefixCache.cpp:34
↓ 1 callersMethodprefetch
mllm/engine/prefix_cache/ZenFS.cpp:130
↓ 1 callersMethodprefill
mllm/backends/qnn/aot_rt/PromptProcessor.cpp:112
↓ 1 callersMethodprefillEventEndTimePoint
mllm/models/ARGeneration.cpp:498
↓ 1 callersMethodprefillEventStartTimePoint
mllm/models/ARGeneration.cpp:496
↓ 1 callersMethodprefillText
* Prefill a chunk of new text tokens in streaming setting */
mllm/models/minicpm_o2_6/modeling_chattts.hpp:204
↓ 1 callersFunctionprepareParametersForSHA
* @brief Prepares the parameter file by slicing MHA weights into SHA weights. * * This function takes the original parameter file with MHA weights a
examples/qwen2_qnn_aot/modeling_qwen2_qnn_aot_sha.hpp:657
↓ 1 callersFunctionprepareParametersForSHA
* @brief Prepares the parameter file by slicing MHA weights into SHA weights. * * This function takes the original parameter file with MHA weights a
examples/llama_qnn_aot/modeling_llama_qnn_aot_sha.hpp:657
↓ 1 callersFunctionprepareParametersForSHA
* @brief Prepares the parameter file by slicing MHA weights into SHA weights. * * This function takes the original parameter file with MHA weights a
examples/qwen3_qnn_aot/modeling_qwen_qnn_aot_sha.hpp:670
↓ 1 callersMethodpreprocessRequestForOCR
(payload map[string]interface{})
mllm-cli/pkg/server/handlers.go:47
↓ 1 callersFunctionprint2DVetcors
mllm/preprocessor/audio/Audio.cpp:153
↓ 1 callersMethodprintModelInfo
tools/mllm-llm-benchmark/models/Qwen3_W4A32_KAI.hpp:35
↓ 1 callersMethodprocess
mllm/models/minicpm_o2_6/image_preprocessor_minicpmo.hpp:253
↓ 1 callersMethodprocessAudioData
Process raw audio data and extract mel-spectrogram features
mllm/models/minicpm_o2_6/audio_preprocessor_minicpmo.hpp:302
← previousnext →1,501–1,600 of 5,971, ranked by callers