 Pascal
|
683fa6ba4e
fix: added a normalization step for MathJax-style \[\] and \(\) delimiters (#16599)
|
3 months ago |
 GittyBurstein
|
b22572e97d
sycl : add ARANGE operator (#16362)
|
3 months ago |
 Chenguang Li
|
7a50cf388a
CANN: format code using .clang-format (#15863)
|
3 months ago |
 takasurazeem
|
6f5d924637
common : Update the docs on -t --threads (#16236)
|
3 months ago |
 takuya kodama
|
adc9b60f19
ggml-cpu: replace putenv with setenv for const-correctness (#16573)
|
3 months ago |
 yael-works
|
ee50ee1ead
SYCL: Add GGML_OP_MEAN operator support (#16009)
|
3 months ago |
 Aleksei Nikiforov
|
7adc79c032
gguf-py : add support for endian conversion of BF16 data (#16594)
|
3 months ago |
 safranowith
|
466c1911ab
cpu : add FLOOR, CEIL, ROUND and TRUNC unary operators (#16083)
|
3 months ago |
 lhez
|
0cb7a0683b
opencl: add q8_0 mm support (#16469)
|
3 months ago |
|
lhez
|
d93f8439b0
opencl: fix FA for f32 (#16584)
|
3 months ago |
 Aleksander Grygier
|
f9fb33f263
Add server-driven parameter defaults and syncing (#16515)
|
3 months ago |
 Sam/Samuel
|
f4ce81c45e
metal: optimise `GGML_OP_SUM` (#16559)
|
3 months ago |
 Georgi Gerganov
|
17304cbcc1
server : fix img token logs (#16595)
|
3 months ago |
 Xuan-Son Nguyen
|
3e3cb19f64
llama-quant: add support for mmproj (#16592)
|
3 months ago |
 Julius Tischbein
|
5acd455460
CUDA: Changing the CUDA scheduling strategy to spin (#16585)
|
3 months ago |
|
Georgi Gerganov
|
554fd578a5
server : fix mtmd checkpoints (#16591)
|
3 months ago |
|
Georgi Gerganov
|
fa882fd2b1
metal : avoid using Metal's gpuAddress property (#16576)
|
3 months ago |
 SavicStefan
|
ffa059034c
vulkan: Add ACC_TYPE_VEC2 implementation (#16203)
|
3 months ago |
 Aman Gupta
|
120bf7046d
CUDA + openCL: fix bug in accessing rms_norm->src while doing fusion (#16577)
|
3 months ago |
 Jeff Bolz
|
4258e0cfe7
vulkan: Support FA with K/V in F32 (#16543)
|
3 months ago |
|
Jeff Bolz
|
7ea15bb64c
vulkan: Improve build time for MSVC (#16545)
|
3 months ago |
 Johannes Gäßler
|
9c7185dd28
CUDA: enable FA for FP32 KV cache (#16546)
|
3 months ago |
|
Aman Gupta
|
1ee9d0b415
CUDA: use fastdiv + ggml_cuda_mad for mmvf (#16557)
|
3 months ago |
|
Aman Gupta
|
48e2fa9fb7
CUDA: add fp kernel for larger batch size MoE (#16512)
|
3 months ago |
 Anav Prasad
|
5b6913c47b
cuda : remove legacy copy-op pointer indirection code (#16485)
|
3 months ago |
|
Georgi Gerganov
|
bc07349a7f
server : dynamic token limit for prompt cache (#16560)
|
3 months ago |
|
Georgi Gerganov
|
e60f241eac
metal : FA support F32 K and V and head size = 32 (#16531)
|
3 months ago |
|
Georgi Gerganov
|
e38b7c6e9e
graph : support cacheless embeddings with FA and iSWA (#16528)
|
3 months ago |
|
lhez
|
5016b72862
opencl: fix build targeting CL 2 (#16554)
|
3 months ago |
|
Johannes Gäßler
|
7049736b2d
CUDA: fix numerical issues in tile FA kernel (#16540)
|
3 months ago |
