Georgi Gerganov
|
0fd7ca7a21
authors : update (#12271)
|
10 месяцев назад |
Jason C.H
|
6fefc05a7a
ggml-backend : make path_str compatible with C++20 (#12269)
|
10 месяцев назад |
Georgi Gerganov
|
7ab364390f
server : infill gen ends on new line (#12254)
|
10 месяцев назад |
Daniel Bevenius
|
7c7f3b7f43
ggml : skip intermediate .air file when compiling .metallib (#12247)
|
10 месяцев назад |
Georgi Gerganov
|
102ac1891d
sync : ggml
|
10 месяцев назад |
vmobilis
|
d6ae2fa061
ggml : ggml_compute_forward_concat() for arbitrary tensor type (ggml/1118)
|
10 месяцев назад |
Rémy O
|
68d0027f3d
ggml-cpu: faster AVX2 variant for IQ1_M (#12216)
|
10 месяцев назад |
Georgi Gerganov
|
ea002810a2
ci : fix save-load test invocations (#12245)
|
10 месяцев назад |
Sigbjørn Skjæret
|
8fad3c7a7c
server : Log original chat template parsing error (#12233)
|
10 месяцев назад |
Olivier Chafik
|
7cf64f6bee
sync: minja - support QwQ-32B (#12235)
|
10 месяцев назад |
BB-fat
|
5e2d57b2b2
metal : simplify kernel arguments using a struct (#3229) (#12194)
|
10 месяцев назад |
David Huang
|
f1648e91cf
HIP: fix rocWMMA build flags under Windows (#12230)
|
10 месяцев назад |
Daniel Bevenius
|
d6c95b0740
metal : fix default.metallib build (#12224)
|
10 месяцев назад |
lhez
|
d76a86d967
opencl: Noncontiguous `norm`, `rms_norm`, disable `fp16` for some ops (#12217)
|
10 месяцев назад |
xiaofei
|
776f9e59cc
cmake : fix undefined reference errors for std::filesystem in ggml (#12092) (#12094)
|
10 месяцев назад |
Lucas Moura Belo
|
3d652bfddf
readme : update bindings (#12229)
|
10 месяцев назад |
Johannes Gäßler
|
5220a16d18
CUDA: fix FA logic for PTX 7.0 and CC >= 7.5 (#12222)
|
10 месяцев назад |
David Huang
|
3ffbbd5ce1
HIP: rocWMMA documentation and enabling in workflow builds (#12179)
|
10 месяцев назад |
Olivier Chafik
|
42994048a3
update function-calling.md w/ template override for functionary-small-v3.2 (#12214)
|
10 месяцев назад |
Aaron Teo
|
e9b2f84f14
llava: add big-endian conversion for image encoder (#12218)
|
10 месяцев назад |
uvos
|
e721c05c93
HIP/CUDA: set the paramerter value in maintain_cuda_graph instead of replaceing it. (#12209)
|
10 месяцев назад |
Han Yin
|
57b6abf85a
android : fix KV cache log message condition (#12212)
|
10 месяцев назад |
Henry Linjamäki
|
94bb63e4f0
opencl : fix buffer alignment (#12197)
|
10 месяцев назад |
Henry Linjamäki
|
f79243992c
opencl : fix `ulong` kernel args were set from `int` variables (#12174)
|
10 месяцев назад |
simon886212
|
ed4ce0dda2
opencl : fix profile-related errors (#12095)
|
10 месяцев назад |
Rémy O
|
07d1572347
ggml-cpu: Faster IQ1 mul_mat_vec on AVX2 using BMI2 instructions (#12154)
|
10 месяцев назад |
Akarshan Biswas
|
5e43f104cc
SYCL: Disable f16 Unary OPs as not supported by the kernels (#12201)
|
10 месяцев назад |
Plamen Minev
|
16e4b22c5e
ggml : fix GGMLMetalClass ODR (#12200)
|
10 месяцев назад |
Daniel Bevenius
|
074c4fd39d
ci : add fetch-depth to xcframework upload (#12195)
|
10 месяцев назад |
Olivier Chafik
|
669912d9a5
`tool-call`: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars (#12034)
|
10 месяцев назад |