git.djapps.eu Git - pkg/ggml/sources/ggml/shortlog

]> git.djapps.eu Git - pkg/ggml/sources/ggml/shortlog

overview / pkg / ggml / sources / ggml / shortlog

2025-05-01	SXX	ggml: move fp16/bf16 conversion optimizations to CPU...	commit \| commitdiff \| tree
2025-05-01	Xuan-Son Nguyen	clip : fix pixtral on some GPU backends (llama/13097)	commit \| commitdiff \| tree
2025-05-01	Neo Zhang Jianyu	change the reorder tensor from init to execute OP ...	commit \| commitdiff \| tree
2025-05-01	Radoslav Gerganov	rpc : do not wait for response when sending RPC_CMD_SET...	commit \| commitdiff \| tree
2025-04-30	Diego Devesa	ggml : fix ggml_gallocr_ptr type (#1205)	commit \| commitdiff \| tree
2025-04-30	Georgi Gerganov	media : rm logos (#1203)	commit \| commitdiff \| tree
2025-04-29	Georgi Gerganov	sync : whisper.cpp	commit \| commitdiff \| tree
2025-04-29	Georgi Gerganov	cuda : fix unused variable compile warning (whisper/0)	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	opencl : remove obsolete files (skip) (#1200)	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	sync : llama.cpp upstream/0.0.1982	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	metal : add memory pool for temp allocs (llama/12850)	commit \| commitdiff \| tree
2025-04-24	lhez	opencl: split ggml-opencl.cl into multiple files and...	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	ggml : fix trailing whitespaces (llama/0)	commit \| commitdiff \| tree
2025-04-24	Johannes Gäßler	CUDA: use switch statements in constexpr functions...	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	metal : fix floating-point range of attention scores...	commit \| commitdiff \| tree
2025-04-24	Eve	vulkan: matmul gcn tuning (llama/13016)	commit \| commitdiff \| tree
2025-04-24	Johannes Gäßler	CUDA: noncont MMVQ + batched bs1 MUL_MAT_ID (llama...	commit \| commitdiff \| tree
2025-04-24	Diego Devesa	ggml : add SSE 4.2 and x64 base variant for CPUs withou...	commit \| commitdiff \| tree
2025-04-24	Akarshan Biswas	SYCL: Add non-contiguous support in ROPE (llama/12993)	commit \| commitdiff \| tree
2025-04-24	Jeff Bolz	vulkan: support noncontiguous rms_norm (llama/13031)	commit \| commitdiff \| tree
2025-04-24	Jeffrey Morgan	metal: add neg operator (llama/13029)	commit \| commitdiff \| tree
2025-04-24	Akarshan Biswas	SYCL: Refactor and enable FP16 in binary broadcast...	commit \| commitdiff \| tree
2025-04-24	Radoslav Gerganov	rpc : add RPC_CMD_HELLO (llama/12955)	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	graph : make FA compatible with MLA + add initial Metal...	commit \| commitdiff \| tree
2025-04-24	Alan Gray	ggml: Re-enable CUDA graphs in presence of CONT and...	commit \| commitdiff \| tree
2025-04-24	hipudding	CANN: Add support for async operator submission (llama...	commit \| commitdiff \| tree
2025-04-24	kimminsu	opencl: fix incorrect local_size index in profiling...	commit \| commitdiff \| tree
2025-04-24	Jeff Bolz	vulkan: enable coopmat2 FA gqa and split_k optimization...	commit \| commitdiff \| tree
2025-04-24	Chenguang Li	CANN: Add 310P operator support check (llama/12962)	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	metal : add FA-vec kernels for head size 96 (llama...	commit \| commitdiff \| tree
2025-04-24	hipudding	CANN: Add x86 build ci (llama/12950)	commit \| commitdiff \| tree
2025-04-24	David Huang	CUDA/HIP: Share the same unified memory allocation...	commit \| commitdiff \| tree
2025-04-24	Akarshan Biswas	SYCL: Add ROPE vision kernel (llama/12887)	commit \| commitdiff \| tree
2025-04-24	Srihari-mcw	ggml : Add AVX512 implementation of GEMM - Q4_Kx8 ...	commit \| commitdiff \| tree
2025-04-24	Chenguang Li	CANN: Opt ROPE optimization (llama/12865)	commit \| commitdiff \| tree
2025-04-24	Xinpeng Dou	CANN: Optimize CANN buffer pool memory management ...	commit \| commitdiff \| tree
2025-04-24	Akarshan Biswas	SYCL: Fix im2col (llama/12910)	commit \| commitdiff \| tree
2025-04-24	Radoslav Gerganov	rpc : use ggml_context_ptr (llama/12938)	commit \| commitdiff \| tree
2025-04-24	Georgi Gerganov	scripts : update sync-llama-am.sh	commit \| commitdiff \| tree
2025-04-19	Leonard Mosescu	tests : Fix a few small Windows / MSVC build issues...	commit \| commitdiff \| tree
2025-04-17	Acly	ggml : Depthwise 2D convolution (#1152)	commit \| commitdiff \| tree
2025-04-14	Georgi Gerganov	sync : llama.cpp	commit \| commitdiff \| tree
2025-04-14	SXX	ggml: use _mm[512/256]_dpbusd[_avx]_epi32 to directly...	commit \| commitdiff \| tree
2025-04-14	Alan Gray	ggml: disable CUDA graphs for unsupported DUP and CONT...	commit \| commitdiff \| tree
2025-04-14	Jeff Bolz	vulkan: use aligned loads for flash attention mask...	commit \| commitdiff \| tree
2025-04-14	Ewan Crawford	sycl: Support sycl_ext_oneapi_limited_graph (llama...	commit \| commitdiff \| tree
2025-04-14	Akarshan Biswas	SYCL: Add fp16 type support to unary op kernels (llama...	commit \| commitdiff \| tree
2025-04-14	Aaron Teo	ggml: fix compilation error s390x (llama/12848)	commit \| commitdiff \| tree
2025-04-14	Georgi Gerganov	tests : fix init order (llama/0)	commit \| commitdiff \| tree
2025-04-11	cmdr2	cpu: fix cpu backend's supports-op for GET_ROWS_BACK...	commit \| commitdiff \| tree
2025-04-10	Georgi Gerganov	sync : fix (skip) (#0)	commit \| commitdiff \| tree
2025-04-10	Georgi Gerganov	sync : llama.cpp	commit \| commitdiff \| tree
2025-04-10	Chenguang Li	CANN: Support more ops (llama/12841)	commit \| commitdiff \| tree
2025-04-10	Prajwal B Mehendarkar	Fixes #12823 (llama/12830)	commit \| commitdiff \| tree
2025-04-10	Piotr Kubaj	ggml-cpu-impl.h: do not redefine bool on POWER9 (llama...	commit \| commitdiff \| tree
2025-04-10	Piotr Kubaj	ggml-impl.h: fix build on POWER9 (llama/12855)	commit \| commitdiff \| tree
2025-04-10	Chenguang Li	CANN: Support Opt CONV_TRANSPOSE_1D and ELU (llama...	commit \| commitdiff \| tree
2025-04-10	Jeff Bolz	vulkan: In coopmat2 mmq, load q4_k/q5_k scales through...	commit \| commitdiff \| tree
2025-04-10	Jeff Bolz	vulkan: Use fp16 for the flash attention P*V multiplica...	commit \| commitdiff \| tree
2025-04-10	Sigbjørn Skjæret	cuda : add f32 to bf16 copy op (llama/12806)	commit \| commitdiff \| tree
2025-04-10	Georgi Gerganov	llama : fix FA when KV cache is not used (i.e. embeddin...	commit \| commitdiff \| tree
2025-04-10	cmdr2	ggml: don't include arm_neon.h when using CUDA 12 with...	commit \| commitdiff \| tree
2025-04-09	Diego Devesa	ggml : add bilinear upscale support (#1185)	commit \| commitdiff \| tree
2025-04-09	Diego Devesa	ggml : add more generic custom op, remove deprecated...	commit \| commitdiff \| tree
2025-04-08	Georgi Gerganov	sync : llama.cpp	commit \| commitdiff \| tree
2025-04-08	Neo Zhang Jianyu	Revert "sycl:remove redundant memcopy in function ggml_...	commit \| commitdiff \| tree
2025-04-08	lhez	opencl: better identify Adreno GPU (llama/12760)	commit \| commitdiff \| tree
2025-04-08	Georgi Gerganov	cuda : fix HIP and MUSA BF16 (llama/0)	commit \| commitdiff \| tree
2025-04-08	zhouwg	sycl: remove redundant memcopy in function ggml_backend...	commit \| commitdiff \| tree
2025-04-08	zhouwg	CANN: fix typo in ggml-cann (llama/12733)	commit \| commitdiff \| tree
2025-04-08	hipudding	CANN: Refactor to reduce duplicate code (llama/12731)	commit \| commitdiff \| tree
2025-04-08	R0CKSTAR	musa: fix compilation warnings in mp_22/31 (llama/12780)	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: fix NaN issue in flash attention shader (llama...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: Use unclamped loads for flash attention mask...	commit \| commitdiff \| tree
2025-04-08	0cc4m	Vulkan: Tune Vulkan mmq int dot shader for performance...	commit \| commitdiff \| tree
2025-04-08	Nicolò Scipione	sycl: allow ggml-sycl configuration and compilation...	commit \| commitdiff \| tree
2025-04-08	Ronny Brendel	cmake: fix ggml-shaders-gen compiler paths containing...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: Hybrid waitForFences/getFenceStatus to reduce...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: set cmake minimum and project name in vulkan...	commit \| commitdiff \| tree
2025-04-08	Gaurav Garg	CUDA: Prefer vector flash decoding kernel for Gemma...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: Fix missing cmake logic for dot product extensi...	commit \| commitdiff \| tree
2025-04-08	a3sh	fix MUSA compiler warning (llama/12704)	commit \| commitdiff \| tree
2025-04-08	Chenguang Li	CANN: Support operator SIN COS ARGMAX (llama/12709)	commit \| commitdiff \| tree
2025-04-08	Alan Gray	Simplify and improve CUDA graphs through use of indirec...	commit \| commitdiff \| tree
2025-04-08	hipudding	CANN: Fix failed test cases (llama/12708)	commit \| commitdiff \| tree
2025-04-08	lhez	opencl: use `max_alloc_size` in backend ctx instead...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: Implement split_k for coopmat2 flash attention...	commit \| commitdiff \| tree
2025-04-08	bandoti	cmake: remove caching from vulkan coopmat checks (llama...	commit \| commitdiff \| tree
2025-04-08	Jeff Bolz	vulkan: Implement grouped query attention in the coopma...	commit \| commitdiff \| tree
2025-04-08	0cc4m	Vulkan: Fix mmq int dot float cache size (llama/12722)	commit \| commitdiff \| tree
2025-04-08	Diego Devesa	llama : add option to override model tensor buffers...	commit \| commitdiff \| tree
2025-04-07	Georgi Gerganov	ggml : simplify Arm fp16 CPU logic (#1177)	commit \| commitdiff \| tree
2025-04-04	Sigbjørn Skjæret	CUDA: don't convert BF16 weights to FP32 (#1174)	commit \| commitdiff \| tree
2025-04-03	Georgi Gerganov	sync : whisper.cpp upstream/0.0.1898	commit \| commitdiff \| tree
2025-04-02	cmdr2	cpu: move all the operators into a separate c++ file...	commit \| commitdiff \| tree
2025-04-02	Georgi Gerganov	sync : llama.cpp	commit \| commitdiff \| tree
2025-04-02	Chenguang Li	get_rows and dup optimization (llama/12671)	commit \| commitdiff \| tree
2025-04-02	Junil Kim	opencl : fix memory allocation size (llama/12649)	commit \| commitdiff \| tree
2025-04-02	Georgi Gerganov	metal : use F32 prec in FA kernels (llama/12688)	commit \| commitdiff \| tree
2025-04-02	R0CKSTAR	Fix clang warning in gguf_check_reserved_keys (llama...	commit \| commitdiff \| tree
next

Packaging of ggml-org/ggml

RSS Atom