Tags
Tags give the ability to mark specific points in history as being important
master-98ed165
98ed1655
·
OpenCL: Add release memory (#1741)
·
Jun 09, 2023
master-0bf7cf1
0bf7cf1b
·
Revert "ggml : load data into int8x16x4_t using vld4q_s8 on arm64 (#1738)"
·
Jun 08, 2023
master-8432d4d
8432d4d9
·
ggml : load data into int8x16x4_t using vld4q_s8 on arm64 (#1738)
·
Jun 08, 2023
master-b50b570
b50b570e
·
ggml : fix fprintf warnings (#1720)
·
Jun 08, 2023
master-0035858
00358582
·
k-quants : add missing compile definition to CMakeLists (#1748)
·
Jun 08, 2023
master-5c64a09
5c64a095
·
k-quants : allow to optionally disable at compile time (#1734)
·
Jun 07, 2023
master-35a8491
35a84916
·
main: add the possibility to open the prompt cache read-only (#1640)
·
Jun 06, 2023
master-2d7bf11
2d7bf110
·
llama : fix vram_scratch var
·
Jun 06, 2023
master-17366df
17366df8
·
Multi GPU support, CUDA refactor, CUDA scratch buffer (#1703)
·
Jun 06, 2023
master-44f906e
44f906e8
·
metal : add f16 support
·
Jun 06, 2023
master-d5b111f
d5b111f5
·
Clblast fixes + enhancements to save VRAM and offload more layers (#1675)
·
Jun 06, 2023
master-2d43387
2d43387d
·
ggml : fix builds, add ggml-quants-k.o (close #1712, close #1710)
·
Jun 06, 2023
master-7a74dee
7a74dee6
·
llama : temporary disable Q6_K output quantization (#1711)
·
Jun 06, 2023
master-590250f
590250f7
·
metal : add checks for buffer size (#1706)
·
Jun 06, 2023
master-c2df36d
c2df36d6
·
llama : consistently catch and throw only exceptions deriving from std::exception (#1599)
·
Jun 05, 2023
master-9d0693b
9d0693bc
·
metal : use shared buffers between CPU and GPU (#1696)
·
Jun 05, 2023
master-efe0507
efe05076
·
ggml : fix internal overflow in ggml_time_us on Windows (#1702)
·
Jun 05, 2023
master-e7fe66e
e7fe66e6
·
ci : disable auto tidy (#1705)
·
Jun 05, 2023
master-99009e7
99009e72
·
ggml : add SOTA 2,3,4,5,6 bit k-quantizations (#1684)
·
Jun 05, 2023
master-5220a99
5220a991
·
Increase 3B scratch buffers. (#1698)
·
Jun 05, 2023
1
…
16
17
18
19
20
21
22
23
24
…
38