 slaren
|
95f57bb5d5
ggml : remove ggml_task_type and GGML_PERF (#8017)
|
1 rok temu |
 0cc4m
|
45c0e2e4c1
Refactor Vulkan backend to allow multiple contexts (#7961)
|
1 rok temu |
 k.h.lai
|
557b653dc9
vulkan: detect multiple devices by deviceUUID instead of deviceID (#8022)
|
1 rok temu |
|
0cc4m
|
7c7836d9d4
Vulkan Shader Refactor, Memory Debugging Option (#7947)
|
1 rok temu |
|
slaren
|
f578b86b21
move BLAS to a separate backend (#6210)
|
1 rok temu |
 Georgi Gerganov
|
a9cae48003
tests : add non-cont unary tests (#7857)
|
1 rok temu |
|
k.h.lai
|
73bac2b11d
vulkan: select only one device for single gpu with multiple drivers (#7582)
|
1 rok temu |
|
0cc4m
|
ef52d1d16a
Update Vulkan RoPE implementation (#7818)
|
1 rok temu |
|
slaren
|
da799b4189
vulkan : reuse parent extra for views (#7806)
|
1 rok temu |
|
Georgi Gerganov
|
2b3389677a
ggml : refactor rope norm/neox (#7634)
|
1 rok temu |
|
0cc4m
|
3d7ebf6312
Vulkan Mixture of Experts (MoE) support (#7628)
|
1 rok temu |
|
k.h.lai
|
56411a950f
vulkan: properly initialize vulkan devices for LLAMA_SPLIT_MODE_NONE (#7552)
|
1 rok temu |
|
0cc4m
|
1b1e27cb49
Update vulkan rope implementation to support frequency factors (#7475)
|
1 rok temu |
 liuwei-git
|
201cc11afa
llama : add phi3 128K model support (#7225)
|
1 rok temu |
|
0cc4m
|
f030ec1f7a
Vulkan Embedding Fix (#7360)
|
1 rok temu |
|
0cc4m
|
c1b295eea5
Update and fix Vulkan soft_max and argsort implementations (#7237)
|
1 rok temu |
|
Georgi Gerganov
|
9cb317f77e
ggml : full ALiBi support (#7192)
|
1 rok temu |
|
0cc4m
|
befddd0f15
Vulkan Bugfixes and Improvements (#7084)
|
1 rok temu |
|
Georgi Gerganov
|
9c67c2773d
ggml : add Flash Attention (#5021)
|
1 rok temu |
|
0cc4m
|
ba0c7c70ab
Vulkan k-quant mmq and ggml-backend offload functionality (#6155)
|
1 rok temu |
 compilade
|
557410b8f0
llama : greatly reduce output buffer memory usage (#6122)
|
1 rok temu |
|
slaren
|
2bf8d0f7c4
backend : offload large batches to GPU (#6083)
|
1 rok temu |
 GainLee
|
dc0f612548
ggml:fix finding transfer queue family index error (#6094)
|
1 rok temu |
|
slaren
|
f30ea47a87
llama : add pipeline parallelism support (#6017)
|
1 rok temu |
|
Georgi Gerganov
|
5b09797321
ggml : remove old quantization functions (#5942)
|
1 rok temu |
|
0cc4m
|
61d1c88e15
Vulkan Improvements (#5835)
|
1 rok temu |
 Michael Podvitskiy
|
9fa2627347
ggml : introduce ggml_status (ggml/750)
|
1 rok temu |
 ddpasa
|
c2224f003b
ggml-vulkan: fix VULKAN_CHECK_RESULTS flag, which was previously broken (#5813)
|
1 rok temu |
 Eve
|
317709b2a8
make portability_enumeration_ext apple only (#5757)
|
1 rok temu |
 UEXTM.com
|
5f70671856
Introduce backend GUIDs (ggml/743)
|
1 rok temu |
