commits

SHA	Message	Author	Date
8c6a9b8b	ruby : Follow source tree change (#2580)	KITAITI Makoto <K**o@gm> Committed by: GitHub <ny@g**m>	almost 2 years ago
37c88027	whisper : use backend registry (#0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
9db070a3	ggml/sched : do not skip views in pre-assignments	slaren <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
7fd8d9c2	whisper : adapt to new ggml (wip)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
06e059b8	talk-llama : sync llama.cpp	Georgi Gerganov <g**v@g**m>	almost 2 years ago
c9f49d5f	sync : ggml	Georgi Gerganov <g**v@g**m>	almost 2 years ago
f4c1d7df	ggml : sync resolve (skip) (#0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
5f7e094c	scripts : update sync	Georgi Gerganov <g**v@g**m>	almost 2 years ago
6266a9f9	release : v1.7.2	Georgi Gerganov <g**v@g**m>	almost 2 years ago
339b8e55	Add required ggml-base and backend libs to cmake pkg (llama/10407)	bandoti <1**i@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
5f6d6919	cuda : fix CUDA_FLAGS not being applied (llama/10403)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
8ee76773	sycl : Add option to set the SYCL architecture for all targets (llama/10266)	Romain Biessy <r**y@cm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
45f1f914	vulkan: Optimize soft_max (llama/10301)	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
53589c8f	sycl: Revert MUL_MAT_OP support changes (llama/10385)	Alberto Cabrera Pérez <a**a@cm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
7ac2f17f	cuda : only use native when supported by cmake (llama/10389)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
48862c7b	vulkan: remove use of null initializer (llama/10372)	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
44f7d9f4	metal : fox offset integer overflows in im2col (ggml/1015)	Plamen Minev <p**v@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
d24f981f	sycl: fix example build (#2570)	Stefan Sydow <s**w@he> Committed by: GitHub <ny@g**m>	almost 2 years ago
fd123025	Vulkan: Fix device info output format specifiers (llama/10366)	0cc4m <p**2@le> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
f80bef46	metal : add `GGML_UNARY_OP_ELU` kernel (ggml/1018)	PAB <p**r@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
161b4435	CUDA: fix MMV kernel being used for FP16 src1 (llama/10357)	Johannes Gäßler <j**g@5e> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
ef7fbe1c	CMake: fix typo in comment [no ci] (llama/10360)	Johannes Gäßler <j**g@5e> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
0879d359	llama : only use default buffer types for the KV cache (llama/10358)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
2a444dc5	metal : refactor kernel args into structs (llama/10238)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
45cf1634	ggml : fix undefined reference to 'getcpu' (llama/10354)	FirstTimeEZ <1**Z@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
dcb2922d	CUDA: remove DMMV, consolidate F16 mult mat vec (llama/10318)	Johannes Gäßler <j**g@5e> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
3c5c7511	CMake: default to -arch=native for CUDA build (llama/10320)	Johannes Gäßler <j**g@5e> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
24ad19d0	ggml : fix possible buffer use after free in sched reserve (llama/9930)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
bd574b05	ggml : inttypes.h -> cinttypes (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
7e0eafcb	ggml : adapt AMX to tensor->grad removal (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
75670ae6	ggml : fix compile warnings (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
01d3bd7d	ci : use local ggml in Android build (#2567)	Georgi Gerganov <g**v@gm> Committed by: GitHub <ny@g**m>	almost 2 years ago
bb12cd9b	ggml : tmp workaround for whisper.cpp (skip) (#2565)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
d4fcdf60	llamafile : fix include path (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
c9541741	ggml: new optimization interface (ggml/988)	Johannes Gäßler <j**g@5e> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
1bebb1a1	vulkan: Optimize some mat-vec mul quant shaders (llama/10296)	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
ee437cde	ggml : optimize Q4_0 into Q4_0_X_Y repack (llama/10324)	Dan Johansson <d**n@am> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
6a55015d	ggml : remove duplicated sources from the last sync (ggml/1017)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
c1506d38	Make updates to fix issues with clang-cl builds while using AVX512 flags (lla...	Srihari-mcw <9**w@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
401fbea3	sync : leftovers (ggml/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
44d1cbdf	cmake : restore CMakeLists.txt (llama/10256)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
7e86030d	ggml : fix some build issues	slaren <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
f02b40bc	update : readme	Georgi Gerganov <g**v@g**m>	almost 2 years ago
83ac2842	scripts : fix sync path	Georgi Gerganov <g**v@g**m>	almost 2 years ago
c4e95fb7	whisper.swiftui : switch Mac dest to Mac (Designed for iPad) (#2562)	Jhen-Jie Hong <d**r@je> Committed by: GitHub <ny@g**m>	almost 2 years ago
3216efef	AVX BF16 and single scale quant optimizations (llama/10212)	Eve <1**e@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
03af461d	ci : fix building workflow for linux/arm64 container (#2555)	Raiya Araki <8**2@um> Committed by: GitHub <ny@g**m>	almost 2 years ago
c0a9f8ef	whisper : include ggml-cpu.h (#0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
6477b84e	build : fixes	Georgi Gerganov <g**v@g**m>	almost 2 years ago
e23721f3	cmake : fix ppc64 check (#0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
24d70677	talk-llama : sync llama.cpp	Georgi Gerganov <g**v@g**m>	almost 2 years ago
5089ab2d	whisper : fix build (#0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
bdbb9068	sync : ggml	Georgi Gerganov <g**v@g**m>	almost 2 years ago
498ac0dc	scripts : update sync	Georgi Gerganov <g**v@g**m>	almost 2 years ago
2c0484eb	sycl: Use syclcompat::dp4a (llama/10267)	Romain Biessy <r**y@cm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
3298916e	backend cpu: add online flow for aarch64 Q4_0 GEMV/GEMM kernels (llama/9921) Co-authored-by: Diego Devesa <s**h@g**m>	Charles Xu <c**u@am> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
746bf259	ggml : build backends as libraries (llama/10256) Co-authored-by: Georgi Gerganov <g**v@gm>, R0CKSTAR <xe@mm> Signed-off-by: Xiaodong Ye <xe@m**m>	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
f19463ec	ruby : extend API (#2551)	KITAITI Makoto <K**o@gm> Committed by: GitHub <ny@g**m>	almost 2 years ago
5f8a086e	whisper.swiftui : add model download list & bench methods (#2546)	Jhen-Jie Hong <d**r@je> Committed by: GitHub <ny@g**m>	almost 2 years ago
a28d82e3	ruby : fix the instructions (#2548)	Wilson Silva <w**s@gm> Committed by: GitHub <ny@g**m>	almost 2 years ago
5ccca19f	ggml : vulkan logs (#2547)	thewh1teagle <6**e@um> Committed by: GitHub <ny@g**m>	almost 2 years ago
300c07b9	examples : fix ffmpeg v5 build (#2543)	Stefan Sydow <s**n@sl> Committed by: GitHub <ny@g**m>	almost 2 years ago
fa2ebd33	sycl : Fixes to broken builds and test-backend-ops (llama/10257)	Alberto Cabrera Pérez <a**a@cm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
21b01a21	vulkan: Optimize contiguous copies (llama/10254)	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
b54ce5ed	vulkan: Throttle the number of shader compiles during the build step. (llama/...	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
26a31b78	metal : more precise Q*K in FA vec kernel (llama/10247)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
14d13c5f	vulkan: Fix newly added tests for permuted mul_mat and 1D im2col (llama/10226)	Jeff Bolz <j**z@nm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
5e110c2e	metal : reorder write loop in mul mat kernel + style (llama/10231)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
4a9926d5	metal : fix build and some more comments (llama/10229)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
ae3c5642	metal : fix F32 accumulation in FA vec kernel (llama/10232)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
e287a3b6	metal : hide debug messages from normal log	Georgi Gerganov <g**v@g**m>	almost 2 years ago
b8902436	ggml: fix zero division in ‘dne’ calculation in CUDA COUNT_EQUAL operator whe...	SXX <s**6@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
b7b38f7d	ggml : optimize llamafile cpu matrix multiplication for ppc64le (llama/10156) Signed-off-by: Amrita H S <a**s@l**m>	amritahs-ibm <a**s@lm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
9f67aab2	metal : opt-in compile flag for BF16 (llama/10218)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
8f0f785d	metal : improve clarity (minor) (llama/10171)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
d0b83357	metal : optimize FA kernels (llama/10171)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
1550be79	ggml : add ggml-cpu.h to the public headers (llama/10204)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
807f848c	fix q4_0_8_8 format for corrupted tokens issue (llama/10198) Co-authored-by: EC2 Default User <e**r@i**l>	snadampal <8**l@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
42398f13	Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleratio... Co-authored-by: Georgi Gerganov <g**v@gm>, Diego Devesa <sh@gm>, Plamen Minev <pv@gm>, Yuri Khrustalev <yv@um>, Meng, Hengyu <al@1**m>	Zhiyuan Li <l**n@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
31aea563	whisper : fix extra memory usage (#2534) Co-authored-by: Vinith Misra <p**n@g**m>	Vin Misra <v**h@au> Committed by: GitHub <ny@g**m>	almost 2 years ago
31c3482a	metal : add BF16 support (llama/8439)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
50257af6	metal : fix from ptr buffer name (llama/10189)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
d111a098	ggml : adjust is_first_call init value (llama/10193)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
915bcd2c	metal : add quantized FA support (llama/10149)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
f69c8b6f	ggml : fix arch check in bf16_to_fp32 (llama/10164)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
8c9044be	Q6_K AVX improvements (llama/10118)	Eve <1**e@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
5f8e9281	ggml : fix gelu tables initialization (llama/10172)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
25da30bd	ggml : fix q4xx mat mul, increase ggml_aligned_malloc alignment (llama/10167)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
54273410	fix build break on arm64 linux (llama/10166)	snadampal <8**l@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
b06b4c0c	cuda : clear error after changing peer access (llama/10153)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
939d36fb	metal : simplify f16 and f32 dequant kernels (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
1471e411	metal : move dequantize templates to beginning of MSL source (llama/0)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
35949192	CANN: adjust backend registry refactor. (llama/10158)	leo-pony <n**a@om> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
9c817edb	ggml : move CPU backend to a separate file (llama/10144)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
24a0feb5	metal : minor fixup in FA kernel (llama/10143)	Georgi Gerganov <g**v@g**m>	almost 2 years ago
28496ac5	cmake : make it possible linking ggml as external lib (ggml/1003)	Yuri Khrustalev <y**v@um> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
2ab8cce7	llama : add simple-chat example (llama/10124) Co-authored-by: Xuan Son Nguyen <t**t@g**m>	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
b40c255e	llama : use smart pointers for ggml resources (llama/10117)	Diego Devesa <s**h@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
ec3e1644	vulkan : improve ggml_vk_create_buffer error handling (llama/9898)	Shupei Fan <d**n@om> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago
b1c06c09	metal : fix minor string leaks (ggml/1004)	Plamen Minev <p**v@gm> Committed by: Georgi Gerganov <gv@g**m>	almost 2 years ago

← Back to repository

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Commits

GitHub / dharmab/whisper.cpp / commits