 Sigbjørn Skjæret
|
e75ba4c043
gguf-py : add support for chat template jinja files (#14508)
|
пре 6 месеци |
 compilade
|
5d46babdc2
llama : initial Mamba-2 support (#9126)
|
пре 6 месеци |
 Georgi Gerganov
|
e17991c466
sync : ggml
|
пре 6 месеци |
 Daniel Bevenius
|
c46944aa25
ggml : add version function to get lib version (ggml/1286)
|
пре 6 месеци |
 Rotem Dan
|
f3ed38d793
Set RPATH to "@loader_path" / "$ORIGIN" to ensure executables and dynamic libraries search for dependencies in their origin directory. (#14309)
|
пре 6 месеци |
 Aman Gupta
|
55a1c5a5fd
CUDA: add softmax broadcast (#14475)
|
пре 6 месеци |
 Johannes Gäßler
|
12a81af45f
CUDA: broadcasting for FlashAttention mask (#14500)
|
пре 6 месеци |
 Jeff Bolz
|
8875523eb3
vulkan: support softmax/FA batch and broadcast (#14449)
|
пре 6 месеци |
|
Georgi Gerganov
|
ec68e84c32
ggml : support bcast ggml_soft_max_ext, ggml_flash_attn_ext (#14435)
|
пре 6 месеци |
 zhouwg
|
307e79d33d
opencl : fix possible buffer overflow in dump_tensor (#14490)
|
пре 6 месеци |
|
Georgi Gerganov
|
d7f5f4e578
simple-chat : fix context-exceeded condition (#14494)
|
пре 6 месеци |
 Eric Zhang
|
c8a4e470f6
opencl : skip empty nodes on cgraph compute (#14491)
|
пре 6 месеци |
 lhez
|
603e43dc91
opencl : update upscale to support align corners (#14488)
|
пре 6 месеци |
|
Sigbjørn Skjæret
|
611ba4b264
ci : add OpenCL to labeler workflow (#14496)
|
пре 6 месеци |
|
Eric Zhang
|
85841e121d
github : add OpenCL backend to issue templates (#14492)
|
пре 6 месеци |
 Björn Ganster
|
68b3cd6514
ggml : Callback before abort (#14481)
|
пре 6 месеци |
|
Georgi Gerganov
|
de56944147
ci : disable fast-math for Metal GHA CI (#14478)
|
пре 6 месеци |
 Grzegorz Grasza
|
1b2aaf28ac
Add Vulkan images to docker.md (#14472)
|
пре 6 месеци |
 Chenguang Li
|
343b6e94b6
CANN: update aclnnGroupedMatmulV2 to aclnnGroupedMatmulV3 (#14411)
|
пре 6 месеци |
|
Jeff Bolz
|
6a746cf9c4
vulkan: Split large mul_mat_id to fit in shared memory (#14451)
|
пре 6 месеци |
|
Sigbjørn Skjæret
|
eff5e45443
add GELU_ERF (#14455)
|
пре 6 месеци |
|
Georgi Gerganov
|
a6a47958a1
ggml : remove trailing whitespace (#0)
|
пре 6 месеци |
|
Georgi Gerganov
|
f61c05d4b1
sync : ggml
|
пре 6 месеци |
 Acly
|
431b2c24f3
ggml-cpu : "align corners" for bilinear upscale/downscale (ggml/1285)
|
пре 6 месеци |
|
Daniel Bevenius
|
497be7c01d
ggml-quants : rename best_mad to best_error (ggml/1283)
|
пре 6 месеци |
|
lhez
|
79b33b2317
opencl : add GEGLU, REGLU, SWIGLU (#14456)
|
пре 6 месеци |
|
Aman Gupta
|
0a5a3b5cdf
Add Conv2d for CPU (#14388)
|
пре 6 месеци |
|
Georgi Gerganov
|
745f11fed0
memory : correctly handle failure in apply() (#14438)
|
пре 6 месеци |
|
Georgi Gerganov
|
5dd942de59
metal : disable fast-math for some cpy kernels (#14460)
|
пре 6 месеци |
 Romain Biessy
|
a7417f5594
ggml-cpu: sycl: Re-enable exp f16 (#14462)
|
пре 6 месеци |
