Georgi Gerganov
|
cf658adc83
llm : add Falcon support (#2717)
|
2 years ago |
Georgi Gerganov
|
ef3f333d37
ggml : sync latest (SAM + SD operators, CUDA alibi) (#2709)
|
2 years ago |
Georgi Gerganov
|
6381d4e110
gguf : new file format with flexible meta data (beta) (#2398)
|
2 years ago |
slaren
|
9e232f0234
ggml : move all type info to ggml_type_traits (#2663)
|
2 years ago |
Georgi Gerganov
|
93356bdb7a
ggml : mul mat tweaks (#2372)
|
2 years ago |
Georgi Gerganov
|
60baff7c85
ggml : pad result of ggml_nbytes()
|
2 years ago |
Georgi Gerganov
|
9082b5dfbf
ggml : change params pointer (style change) (#2539)
|
2 years ago |
Georgi Gerganov
|
99d29c0094
ggml : sync (custom ops) (#2537)
|
2 years ago |
slaren
|
a113689571
ggml : add graph tensor allocator (#2411)
|
2 years ago |
slaren
|
b5472ea0ad
ggml : fix assert in ggml_set_unary_op (#2410)
|
2 years ago |
slaren
|
5488fb789e
ggml : allocate graphs in a context (#2392)
|
2 years ago |
slaren
|
07aaa0f63f
ggml : fix ggml_flash_attn to use op_params (#2387)
|
2 years ago |
Jiahao Li
|
875086bdb9
ggml : relax contiguous constraints in activation function (#2371)
|
2 years ago |
slaren
|
da1889834a
ggml : improve graph build time via hash table lookup (#2329)
|
2 years ago |
slaren
|
41c674161f
make rms_norm_eps a parameter (#2374)
|
2 years ago |
Georgi Gerganov
|
5b2b2dc6ae
ggml : sync (unary ops refactor, static-correctness) (#2370)
|
2 years ago |
slaren
|
3602ac4255
fix n_tasks (#2342)
|
2 years ago |
slaren
|
95a6c595e7
ggml: move op parameters from tensors to ggml_tensor::op_params (#2333)
|
2 years ago |
Georgi Gerganov
|
0db14fef06
ggml : fix the rope fix (513f8619535a64fa9ace808cdcbcf66211535f5c)
|
2 years ago |
Georgi Gerganov
|
513f861953
ggml : fix rope args order + assert (#2054)
|
2 years ago |
Qingyou Meng
|
672dda10e4
ggml : fixed runtime bugs and compile errors related to GGML_PERF and GGML_DEBUG (#2219)
|
2 years ago |
Xiao-Yong Jin
|
6e7cca4047
llama : add custom RoPE (#2054)
|
2 years ago |
Evan Miller
|
e8035f141e
ggml : fix static_assert with older compilers #2024 (#2218)
|
2 years ago |
Georgi Gerganov
|
697966680b
ggml : sync (ggml_conv_2d, fix mul_mat bug, CUDA GLM rope)
|
2 years ago |
Georgi Gerganov
|
975221e954
ggml : broadcast mul_mat + conv batch support (#2199)
|
2 years ago |
Georgi Gerganov
|
4523d10d0c
ggml : add ggml_pool_1d and ggml_pool_2d
|
2 years ago |
Georgi Gerganov
|
20d7740a9b
ggml : sync (abort callback, mul / add broadcast, fix alibi) (#2183)
|
2 years ago |
Spencer Sutton
|
5bf2a27718
ggml : remove src0 and src1 from ggml_tensor and rename opt to src (#2178)
|
2 years ago |
clyang
|
3bbc1a11f0
ggml : fix buidling with Intel MKL but ask for "cblas.h" issue (#2104) (#2115)
|
2 years ago |
Qingyou Meng
|
1d656d6360
ggml : change ggml_graph_compute() API to not require context (#1999)
|
2 years ago |