jacekpoplawski
|
a12363bbf0
convert : text-only support for GLM-4.1V-9B-Thinking (#14823)
|
пре 6 месеци |
Johannes Gäßler
|
a86f52b285
CUDA: fix overflow in FA, tune performance (#14840)
|
пре 6 месеци |
Johannes Gäßler
|
b284197df4
CUDA: fix compilation with GGML_CUDA_F16 (#14837)
|
пре 6 месеци |
Sigbjørn Skjæret
|
221c0e0c58
ci : correct label refactor->refactoring (#14832)
|
пре 6 месеци |
Johannes Gäßler
|
07a19e27a2
CUDA: fix quantized KV cache + multiple sequences (#14822)
|
пре 6 месеци |
Georgi Gerganov
|
18f3b5ff9e
tests : add non-cont K,V FA tests
|
пре 6 месеци |
l3utterfly
|
7233358d29
memory : handle saving/loading null layers in recurrent memory (#14675)
|
пре 6 месеци |
lixing-star
|
6c88b3bb25
ggml: fix loongarch quantize_row_q8_1 error (#14827)
|
пре 6 месеци |
chen fan
|
14c28dfc50
CANN: weight format to NZ for Ascend310P3 (#14407)
|
пре 6 месеци |
Aman Gupta
|
8c988fa41d
CUDA: add fused rms norm (#14800)
|
пре 6 месеци |
Csaba Kecskemeti
|
acd6cb1c41
ggml : model card yaml tab->2xspace (#14819)
|
пре 6 месеци |
Jeff Bolz
|
84712b6043
vulkan: fix rms_norm_mul to handle broadcasting dim0 (#14817)
|
пре 6 месеци |
Molly Sophia
|
d4d1522b20
llama : add model type detection for rwkv7 7B&14B (#14816)
|
пре 6 месеци |
Ed Addario
|
d1aa0cc5d1
imatrix: add option to display importance score statistics for a given imatrix file (#12718)
|
пре 6 месеци |
stduhpf
|
c8ade30036
Mtmd: add a way to select device for vision encoder (#14236)
|
пре 6 месеци |
Sigbjørn Skjæret
|
e28c0b80c2
cuda : implement bf16 cpy ops and enable bf16 cont (#14763)
|
пре 6 месеци |
lhez
|
8e6f8bc875
opencl: remove unreachable `return` (#14806)
|
пре 6 месеци |
Molly Sophia
|
adef81781a
server : allow setting `--reverse-prompt` arg (#14799)
|
пре 6 месеци |
R0CKSTAR
|
48b86c4fdb
cuda: remove linking to cublasLt (#14790)
|
пре 6 месеци |
Sigbjørn Skjæret
|
38d3af1b73
opencl: fix `im2col` when `KW!=KH` (#14803)
|
пре 6 месеци |
rmatif
|
6c9ee3b17e
opencl: add conv2d kernel (#14403)
|
пре 6 месеци |
Romain Biessy
|
cd465d823c
sycl: Fix im2col (#14797)
|
пре 6 месеци |
Charles Xu
|
922042601b
kleidiai: add support for get_rows (#14676)
|
пре 6 месеци |
Radoslav Gerganov
|
2ba1333b35
docs : fix backends table in README.md (#14796)
|
пре 6 месеци |
Jeff Bolz
|
c2e058f1b4
vulkan/cuda: Fix im2col when KW!=KH (#14789)
|
пре 6 месеци |
Molly Sophia
|
c82d48ec23
llama : fix `--reverse-prompt` crashing issue (#14794)
|
пре 6 месеци |
IsaacDynamo
|
b4efd77f8a
server : add parse_special option to /tokenize endpoint (#14783)
|
пре 6 месеци |
Aman Gupta
|
2be60cbc27
docs : fix link for tools/perplexity in README.md (#14780)
|
пре 6 месеци |
rspOverflow
|
b526ad2668
Documentation: Further revisions to the Vulkan section in build.md (#14785)
|
пре 6 месеци |
Aman Gupta
|
938b785764
Clang-format: local files first + fix BinPacking (#14779)
|
пре 6 месеци |