R0CKSTAR
|
d9a63b2f2e
musa: enable freediskspace for docker image build (#12839)
|
9 maanden geleden |
Romain Biessy
|
8ed71242f4
sycl: update documentation to use -no-cnv (#12845)
|
9 maanden geleden |
Plamen Minev
|
381603a775
ci: detach common from the library (#12827)
|
9 maanden geleden |
Xuan-Son Nguyen
|
65a69e6e1b
clip : do not print ftype (#12832)
|
9 maanden geleden |
Georgi Gerganov
|
47277d6d1d
readme : add rpc backend (#12842)
|
9 maanden geleden |
Chenguang Li
|
6e1c4cebdb
CANN: Support Opt CONV_TRANSPOSE_1D and ELU (#12786)
|
9 maanden geleden |
Jeff Bolz
|
0090950f67
vulkan: In coopmat2 mmq, load q4_k/q5_k scales through shared memory (#12833)
|
9 maanden geleden |
Jeff Bolz
|
7ecd780b1a
vulkan: Use fp16 for the flash attention P*V multiplication (#12783)
|
9 maanden geleden |
Sigbjørn Skjæret
|
7538246e7c
cuda : add f32 to bf16 copy op (#12806)
|
9 maanden geleden |
Matt Clayton
|
b32efad2bc
llava: improve clip_ctx destructor to not memleak load_image_size (#12834)
|
9 maanden geleden |
Georgi Gerganov
|
a19b5cef16
llama : fix FA when KV cache is not used (i.e. embeddings) (#12825)
|
9 maanden geleden |
Xuan-Son Nguyen
|
78a1ba0a4f
server : fix thread.join() on exit (#12831)
|
9 maanden geleden |
dm4
|
2dabf759e7
llava: add more helper functions to check projector types in clip context (#12824)
|
9 maanden geleden |
Prajwal B Mehendarkar
|
1d343b4069
arg : Including limits file on AIX (#12822)
|
9 maanden geleden |
characharm
|
8ca6e1c3a4
server : webui : Improve Chat Input with Auto-Sizing Textarea (#12785)
|
9 maanden geleden |
Neo Zhang Jianyu
|
656babd6c2
Revert "sycl:remove redundant memcopy in function ggml_backend_sycl_buffer_set_tensor" (#12812)
|
9 maanden geleden |
compilade
|
a226bc7a9a
gguf-py : support lazy tensor splitting (#12809)
|
9 maanden geleden |
Xuan-Son Nguyen
|
1466621e73
llama : Support llama 4 text-only (#12791)
|
9 maanden geleden |
lhez
|
82974011f3
opencl: better identify Adreno GPU (#12760)
|
9 maanden geleden |
stduhpf
|
4ccea213bc
hellaswag: display estimated score confidence interval (#12797)
|
9 maanden geleden |
Georgi Gerganov
|
1a1ab7e7a4
cuda : fix HIP and MUSA BF16 (#0)
|
9 maanden geleden |
Georgi Gerganov
|
a4e46e28f9
sync : ggml
|
9 maanden geleden |
Georgi Gerganov
|
ff067dbcb9
ggml : simplify Arm fp16 CPU logic (ggml/1177)
|
9 maanden geleden |
Sigbjørn Skjæret
|
36ca8b3628
CUDA: don't convert BF16 weights to FP32 (ggml/1174)
|
9 maanden geleden |
cmdr2
|
995083e4ed
cpu: move all the operators into a separate c++ file (except mul_mat) (ggml/1167)
|
9 maanden geleden |
zhouwg
|
518a01480e
sycl: remove redundant memcopy in function ggml_backend_sycl_buffer_set_tensor (#12734)
|
9 maanden geleden |
Xuan-Son Nguyen
|
e391d3ee8d
ci : no curl on ggml-ci (#12796)
|
9 maanden geleden |
Xuan-Son Nguyen
|
bd3f59f812
cmake : enable curl by default (#12761)
|
9 maanden geleden |
zhouwg
|
52b3d71f12
CANN: fix typo in ggml-cann (#12733)
|
9 maanden geleden |
hipudding
|
d0d5b2232b
CANN: Refactor to reduce duplicate code (#12731)
|
9 maanden geleden |