Spaces:
Running
Running
Commit History
ggml : add mmla kernels for quantized GEMM (llama/4966)
0d50a29
unverified
snadampal
commited on
metal : use autoreleasepool to avoid memory leaks (llama/5437)
c276f12
unverified
ggml-alloc : v3 (ggml/727)
5cffd6f
unverified
slaren
commited on
examples : added audio_ctx argument to main and server (#1857)
469988b
unverified
metal : option to embed MSL source into compiled binary (#1842)
a46b62a
unverified
Didzis Gosko
commited on
examples : initialize context params properly (#1852)
3443ee7
unverified
talk-llama : sync llama.cpp
e6d6e1d
unverified
sync : ggml
94800c5
unverified
src : relocate new backend sources
44cd2d4
unverified
ggml : fix `error C2078: too many initializers` for MSVC ARM64 (llama/5404)
8ebb36c
unverified
Michael Podvitskiy
commited on
CUDA: more warps for mmvq on NVIDIA (llama/5394)
7ab774c
unverified
CUDA: fixed mmvq kernel for bs 2,3,4 and -sm row (llama/5386)
3ff7660
unverified
Basic Vulkan Multi-GPU implementation (llama/5321)
5d130aa
unverified
CUDA: mul_mat_vec_q max. batch size 8 -> 4 (llama/5370)
7aa3216
unverified
Slight quantization improvement for Q4_K and Q5_K (llama/5361)
e3cd020
unverified
CUDA: mul_mat_vec_q for batch sizes > 1 (llama/5351)
ae45b38
unverified
ggml : make use of ggml-quants.h possible in C++ code (llama/5338)
963ade6
unverified
ggml : avoid duplicating function calls using MIN/MAX macros (llama/5325)
9bb2b0a
unverified
iq2_xxs: tune quantization (llama/5320)
11e5f6b
unverified
cuda : fix LLAMA_CUDA_F16 (llama/5262)
5fd8fb7
unverified
slaren
commited on
metal : add im2col F32 dst support (llama/5132)
26aec77
unverified
llava : add MobileVLM support (llama/5132)
f17a416
unverified
JidongZhang-THU
slaren
commited on
ggml : limit n_threads to the max n_tasks (llama/5238)
2645c33
unverified
slaren
commited on
kompute : llama-bench support and ggml_cpu_has_kompute() (llama/5226)
0c9c434
unverified
ggml : add abort_callback for cpu backend (ggml/725)
a8ea91b
unverified
Michael Podvitskiy
commited on
extra : update sync scripts
d99e873
unverified
server : allow CORS request with authorization headers (#1850)
16a6639
unverified
Valentin Gosu
commited on
whisper : expose CUDA device setting in public API (#1840)
d13ee66
unverified
Didzis Gosko
commited on
make : add macOS deployment target option (#1839)
9c90601
unverified
Didzis Gosko
commited on
talk-llama : stream response (#1121)
2193f2b
unverified
sync : ggml (#0)
fded75b
unverified
ggml : fix IQ3_XXS on Metal (llama/5219)
f066321
unverified
sync : ggml (llama/0)
cdb7964
unverified
SOTA 3-bit quants (llama/5196)
4649943
unverified
ggml alloc: Fix for null dereference on alloc failure (llama/5200)
8181686
unverified
Paul Tsochantaris
commited on
Nomic Vulkan backend (llama/4456)
f5fd92d
unverified
ggml : add max buffer sizes to opencl and metal backends (llama/5181)
3d354d0
unverified
slaren
commited on
metal : free metal objects (llama/5161)
ea7167a
unverified
Paul Tsochantaris
commited on
gguf : fix comparison (ggml/715)
80cfca4
unverified
`ggml_cuda_cpy` support for 4d tensors and float16->float32 upcasting (ggml/686)
75d438c
unverified
John Balis
slaren
commited on
