Commit History

ggml: new optimization interface (ggml/988)
dd33ace

JohannesGaessler commited on

ggml : remove duplicated sources from the last sync (ggml/1017)
026d20b

ggerganov commited on

ggml : fix some build issues
c5ba1d1

slaren commited on

sync : leftovers (ggml/0)
0f6c498

ggerganov commited on

cmake : restore CMakeLists.txt (llama/10256)
51a70ff

ggerganov commited on

AVX BF16 and single scale quant optimizations (llama/10212)
e6ffed3

Eve commited on

sycl: Use syclcompat::dp4a (llama/10267)
ce0dc30

Romain Biessy commited on

backend cpu: add online flow for aarch64 Q4_0 GEMV/GEMM kernels (llama/9921)
3541ee8

Charles Xu Diego Devesa commited on

ggml : build backends as libraries (llama/10256)
3dc93f3

Diego Devesa ggerganov R0CKSTAR commited on

scripts : update sync
1741306

ggerganov commited on

release : v1.7.2
414329d
unverified

ggerganov commited on

sycl: fix example build (#2570)
a0dcffc
unverified

Stefan Sydow commited on

ci : use local ggml in Android build (#2567)
72b7501
unverified

ggerganov commited on

ggml : tmp workaround for whisper.cpp (skip) (#2565)
ef26f48
unverified

ggerganov commited on

update : readme
d1fa03c
unverified

ggerganov commited on

scripts : fix sync path
9a2f912
unverified

ggerganov commited on

whisper.swiftui : switch Mac dest to Mac (Designed for iPad) (#2562)
13f2beb
unverified

jhenhong commited on

cmake : fix ppc64 check (#0)
f3c3fca

ggerganov commited on

whisper : include ggml-cpu.h (#0)
cb35171

ggerganov commited on

build : fixes
11d19cb

ggerganov commited on

talk-llama : sync llama.cpp
6bb34fb

ggerganov commited on

whisper : fix build (#0)
dfd316d

ggerganov commited on

sync : ggml
9e83be6

ggerganov commited on

sycl : Fixes to broken builds and test-backend-ops (llama/10257)
9cfb13b

Alberto Cabrera Pérez commited on

vulkan: Optimize contiguous copies (llama/10254)
9974bd6

jeffbolznv commited on

vulkan: Throttle the number of shader compiles during the build step. (llama/10222)
9677a2f

jeffbolznv commited on

metal : more precise Q*K in FA vec kernel (llama/10247)
9160e8f

ggerganov commited on

vulkan: Fix newly added tests for permuted mul_mat and 1D im2col (llama/10226)
76b8073

jeffbolznv commited on

metal : reorder write loop in mul mat kernel + style (llama/10231)
661360d

ggerganov commited on

metal : fix build and some more comments (llama/10229)
93fc215

ggerganov commited on

metal : fix F32 accumulation in FA vec kernel (llama/10232)
228e0b2

ggerganov commited on

metal : hide debug messages from normal log
efefcbb

ggerganov commited on

ggml: fix zero division in ‘dne’ calculation in CUDA COUNT_EQUAL operator when ‘ne’ is small (#10213)
0ecc4d6

sxx-404 commited on

ggml : optimize llamafile cpu matrix multiplication for ppc64le (llama/10156)
18bdb35

amritahs-ibm commited on

metal : opt-in compile flag for BF16 (llama/10218)
5f667d1

ggerganov commited on

metal : improve clarity (minor) (llama/10171)
d68ae7c

ggerganov commited on

metal : optimize FA kernels (llama/10171)
44ff932

ggerganov commited on

ggml : add ggml-cpu.h to the public headers (llama/10204)
936a35f

Diego Devesa commited on

fix q4_0_8_8 format for corrupted tokens issue (llama/10198)
4700b48

snadampal EC2 Default User commited on

Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration (llama/10133)
f58e658

Zhiyuan Li ggerganov Diego Devesa pacominev Yuri Khrustalev Meng, Hengyu commited on

metal : add BF16 support (llama/8439)
847669b

ggerganov commited on

metal : fix from ptr buffer name (llama/10189)
c4d59b9

Diego Devesa commited on

ggml : adjust is_first_call init value (llama/10193)
7e2b09b

ggerganov commited on

metal : add quantized FA support (llama/10149)
f1ea157

ggerganov commited on

ggml : fix arch check in bf16_to_fp32 (llama/10164)
09e4a9b

Diego Devesa commited on

Q6_K AVX improvements (llama/10118)
b4c65b4

Eve commited on

ggml : fix gelu tables initialization (llama/10172)
59dd404

Diego Devesa commited on

ggml : fix q4xx mat mul, increase ggml_aligned_malloc alignment (llama/10167)
ba20d5c

Diego Devesa commited on

fix build break on arm64 linux (llama/10166)
68520c4

snadampal commited on

cuda : clear error after changing peer access (llama/10153)
106cf6f

Diego Devesa commited on