 Georgi Gerganov
|
2b3389677a
ggml : refactor rope norm/neox (#7634)
|
1 year ago |
|
Georgi Gerganov
|
554c247caf
ggml : remove OpenCL (#7735)
|
1 year ago |
|
Georgi Gerganov
|
fb76ec31a9
ggml : fix YARN + add tests + add asserts (#7617)
|
1 year ago |
 Radoslav Gerganov
|
210d99173d
llama-bench : add support for the RPC backend (#7435)
|
1 year ago |
|
Georgi Gerganov
|
72de268bec
ggml : restore ggml_rope_xpos_inplace (ggml/0)
|
1 year ago |
|
Georgi Gerganov
|
0548a4187f
ggml : generalize GGML_OP_CONCAT (#7563)
|
1 year ago |
 Masaya, Kato
|
faa0e6979a
ggml: aarch64: SVE kernels for q8_0_q8_0, q4_0_q8_0 vector dot (#7433)
|
1 year ago |
|
Georgi Gerganov
|
d48c88cbd5
ggml : remove ggml_flash_attn and ggml_flash_ff (#7463)
|
1 year ago |
|
Georgi Gerganov
|
3e5faa8503
cuda : fix rope + add tests (#7452)
|
1 year ago |
 liuwei-git
|
201cc11afa
llama : add phi3 128K model support (#7225)
|
1 year ago |
 Srihari-mcw
|
33c8d50acc
Add provisions for windows support for BF16 code including CMake provision for enabling AVX512_BF16 (#7258)
|
1 year ago |
 slaren
|
344f9126cc
ggml : tag ggml_tensor::backend as deprecated (#7290)
|
1 year ago |
 John Balis
|
48aa8fd1f2
ggml : add `ggml_upscale_ext` (ggml/814)
|
1 year ago |
|
Georgi Gerganov
|
e8a7fd4fb0
metal : support FA without mask + add asserts (#7278)
|
1 year ago |
 Justina Cho
|
f5ef34e428
feat: implemented sigmoid function (ggml/806)
|
1 year ago |
|
Georgi Gerganov
|
9cb317f77e
ggml : full ALiBi support (#7192)
|
1 year ago |
 Justine Tunney
|
3855416027
ggml : introduce bfloat16 support (#6412)
|
1 year ago |
|
Georgi Gerganov
|
9c67c2773d
ggml : add Flash Attention (#5021)
|
1 year ago |
|
slaren
|
017e6999b5
add basic tensor data validation function (#6884)
|
1 year ago |
|
slaren
|
0d56246f4b
ggml : group all experts in a single ggml_mul_mat_id (#6505)
|
1 year ago |
 jiez
|
91c736015b
llama : add gguf_remove_key + remove split meta during quantize (#6591)
|
1 year ago |
 Carolinabanana
|
5dc9dd7152
llama : add Command R Plus support (#6491)
|
1 year ago |
|
slaren
|
08a0c02060
ggml : mul_mat_id use the same tensor for all the experts (#6387)
|
1 year ago |
 compilade
|
557410b8f0
llama : greatly reduce output buffer memory usage (#6122)
|
1 year ago |
 Kawrakow
|
55c1b2a3bb
IQ1_M: 1.75 bpw quantization (#6302)
|
1 year ago |
|
slaren
|
280345968d
cuda : rename build flag to LLAMA_CUDA (#6299)
|
1 year ago |
 Jared Van Bortel
|
94d1b3b411
use _wfopen instead of fopen on Windows (#6248)
|
1 year ago |
 Ondřej Čertík
|
7ce2c77f88
gguf : add support for I64 and F64 arrays (#6062)
|
1 year ago |
|
Georgi Gerganov
|
3fe8d7a17f
ggml : designate enum vals for integer types (#6050)
|
1 year ago |
|
Georgi Gerganov
|
5b09797321
ggml : remove old quantization functions (#5942)
|
1 year ago |
