This website requires JavaScript.
9eb5bfec1a
mtmd : update docs to use llama_model_n_embd_inp (#18999 )
2026-01-22 14:36:32 +01:00
c6926d1d95
server: Reorder methods in server-task.cpp (#19016 )
2026-01-22 22:36:04 +09:00
b70d251076
CUDA: add gqa_ratio 4 for GLM 4.7 flash (#18953 )
2026-01-22 18:51:53 +08:00
5516b9c16a
opencl: add TRI op support (#18979 )
2026-01-21 22:05:54 -08:00
94242a62c0
ggml-zdnn : mark zDNN buffers as non-host (#18967 )
2026-01-22 01:16:21 +01:00
6b99a223e3
ci : update GitHub Actions versions [no ci] (#18935 )
2026-01-22 00:57:18 +01:00
77078e80e5
convert : add Devstral-2 (Ministral3ForCausalLM) arch (#18972 )
2026-01-22 00:55:55 +01:00
c301172f66
jinja: support none|string (#18995 )
2026-01-21 19:24:37 +01:00
3802d3c78f
fix: Use tabular-nums for chat message statistics (#18915 )
2026-01-21 18:46:01 +01:00
9da3dcd753
llama : clarify nemotron-h.cpp comment about RoPE [no ci] (#18997 )
2026-01-21 18:31:34 +01:00
bd544c94a3
vulkan: Remove transfer_ctx, do everything in compute_ctx. (#18945 )
2026-01-21 11:01:40 -06:00
14be5a39b1
common : improve error message when HTTPS is missing but required (#18987 )
2026-01-21 17:58:38 +01:00
fbbf3ad190
server: /v1/responses (partial) (#18486 )
2026-01-22 01:47:23 +09:00
33f890e579
vulkan: support flash attention GQA/split_k with small batches (#18938 )
2026-01-21 10:43:43 -06:00
067b8d7af3
Revert "vulkan: force full subgroups for flash attention to fix intel subgroup crash (#17356 )" (#18831 )
2026-01-22 01:13:43 +09:00
50b7f076a5
vulkan: Use mul_mat_vec_id for small values of n (#18918 )
2026-01-21 09:22:02 -06:00
ad8d85bd94
memory : add llama_memory_hybrid_iswa (#18601 )
2026-01-21 13:30:23 +01:00
12a4a47e6a
Fix GLM 4.7 Lite MoE gating func (#18980 )
2026-01-21 12:35:20 +01:00
37c35f0e1c
gguf: display strerrno when cant load a model (#18884 )
2026-01-21 07:52:46 +01:00
5bd341c9a1
CUDA: Fix builds for older CCCL versions by ifdefing strided_iterator (#18964 )
2026-01-21 02:34:29 +01:00
1c7cf94b22
common, server : use the same User-Agent by default (#18957 )
2026-01-20 18:28:43 +01:00
2c1f199653
cli : fix reasoning responses in CLI (#18961 )
2026-01-20 18:23:25 +01:00
d1e3556481
CUDA: Replace init_offsets kernel with iterators in cub-based argsort (#18930 )
2026-01-20 13:11:01 +01:00
08f3f4a8a3
ggml : cleanup path_str() (#18928 )
2026-01-20 11:42:49 +01:00
271191906c
metal : enable FA for MLA heads (#18950 )
2026-01-20 12:21:28 +02:00
7dee9ff59a
convert : use n_groups instead of hardcoded values in reshape (#18929 )
2026-01-20 06:55:24 +01:00
6df686bee6
server : refactor oai_parser_opt, move it to server_chat_params (#18937 )
2026-01-19 23:28:01 +01:00
1706a6d7c6
convert : support Glm4MoeLite (#18936 )
2026-01-19 16:09:20 -06:00
959ecf7f23
jinja : fix undefined keys and attributes and int/float as bool (#18924 )
2026-01-19 20:29:43 +01:00
4037093c66
ci : run test-jinja -py on high perf [no ci] (#18916 )
2026-01-19 20:29:15 +01:00
18361c579c
server: fix memory reservations in populate_token_probs (#18787 )
2026-01-19 19:13:31 +01:00
365a3e8c31
ggml : add ggml_build_forward_select (#18550 )
2026-01-19 20:03:19 +02:00
3d55846a5c
model-conversion : add BUILD_DIR variable to run-converted-model scripts (#18927 )
2026-01-19 13:12:38 +01:00
287a33017b
llama : Extend fallback, fix fileno for dio file, exclude case that mmap uses dio file (#18887 )
2026-01-18 17:35:57 +01:00
293a1565dc
docs: add linux to index (#18907 )
2026-01-18 05:03:35 -05:00
fe44d35574
tests : add test-jinja -py option for cross-checking (#18906 )
2026-01-18 08:14:27 +01:00
bbcdac0189
jinja : fix object item order (and properly implement dictsort) (#18904 )
2026-01-18 03:40:06 +01:00
d03c45c9c5
jinja : attribute support for join, map and sort (#18883 )
2026-01-18 02:53:01 +01:00
10c98cbdf6
jinja : add missing tojson filter for bool (#18900 )
2026-01-18 01:05:09 +01:00
420960ab92
jinja : fix lexing of float literals with sign (#18901 )
2026-01-18 00:57:51 +01:00
f55b033ae6
jinja: correct member access rule (#18905 )
2026-01-18 00:48:55 +01:00
d1b4757ded
opencl: fix q6_K mv for m=1 (#18893 )
2026-01-17 13:50:32 -08:00
57c0beaed0
ci : add label for jinja changes (#18903 )
2026-01-17 21:52:02 +01:00
2fbde785bc
kv-cache : optimize KQ mask construction (#18842 )
2026-01-17 15:42:42 +02:00
a89002f07b
ggml webgpu: support for backend sampling (#18880 )
2026-01-16 16:12:43 -08:00
388ce82241
ggml : extend ggml_pool_1d + metal (#16429 )
2026-01-16 15:59:56 +01:00
6ba6a3c76f
docs : update ops.md for CANN backend (#18654 )
2026-01-16 20:32:17 +08:00
0802d4cfb3
ggml-blas: hide warnings from included BLAS headers (#18818 )
2026-01-16 06:38:25 -05:00
c945aaaef2
mtmd : Fix ASR for LFM2.5-Audio-1.5B (#18876 )
2026-01-16 11:23:08 +01:00
c15395f73c
common : implement new jinja template engine (#18462 )
2026-01-16 11:22:06 +01:00
aa1dc3770a
Setting mmap and direct_io to false as default in llama-bench.cpp (#18841 )
2026-01-16 09:46:51 +01:00
4ea2eaac01
CANN: Remove unused ggml_cann_get_device function (#18625 )
2026-01-16 08:34:09 +00:00
e20fa27a02
CANN: fix an issue where get_env was not fully renamed (#18796 )
2026-01-16 16:24:04 +08:00
baa4ba0aec
CANN: support gated linear attn (#18653 )
2026-01-16 16:18:49 +08:00
785a710085
OpenCL: add SOLVE_TRI op support (#18846 )
2026-01-15 11:17:17 -08:00
6e7fc8a146
cuda : print less debug logs when disabling cuda graphs (#18868 )
2026-01-15 20:53:01 +02:00
be8e3d9515
context : do not reserve scheduler for warmups (#18867 )
2026-01-15 19:35:57 +02:00
13f1e4a9ca
llama : add adaptive-p sampler (#17927 )
2026-01-15 11:16:29 -06:00
a04c2b06a3
server: improve slots scheduling for n_cmpl (#18789 )
2026-01-15 17:10:28 +01:00
39173bcacb
context : reserve new scheduler when graph topology changes (#18547 )
2026-01-15 16:39:17 +02:00
5c662d21a3
CUDA: fix allignment on register spill for FA (#18815 )
2026-01-15 15:14:50 +01:00
8cc0ba957b
ggml-cpu: optimize ggml_vec_dot_bf16 for Power9 (#18837 )
2026-01-15 15:01:18 +05:30
a7e6ddb8bd
lora: make sure model keep track of associated adapters (#18490 )
2026-01-15 10:24:28 +01:00
2a13180100
model-loader : support bool array sliding window pattern (#18850 )
2026-01-15 10:12:46 +01:00
ec997b4f2b
tests : download models only when running ctest (#18843 )
2026-01-15 09:47:29 +01:00
cff777f226
hexagon: support for OP_CPY, host buffers now optional, hvx-utils refactoring and optimizations (#18822 )
2026-01-14 21:46:12 -08:00
36f0132464
CUDA: Factor out and re-use block_reduce function (#18785 )
2026-01-15 03:44:54 +01:00
d98b548120
Restore clip's cb() to its rightful glory - extract common debugging elements in llama (#17914 )
2026-01-14 20:29:35 +01:00
8fb7175576
model : clean up and fix EXAONE-MoE configuration (#18840 )
2026-01-15 03:38:21 +09:00
516a4ca9b5
refactor : remove libcurl, use OpenSSL when available (#18828 )
2026-01-14 18:02:47 +01:00
3e4bb29666
vulkan: Check maxStorageBufferRange in supports_op (#18709 )
2026-01-14 03:59:05 -06:00
47f9612492
llama-model: fix unfortunate typo (#18832 )
2026-01-14 17:55:15 +08:00
01cbdfd7eb
CUDA : fix typo in clang pragma comment [no ci] (#18830 )
2026-01-14 10:31:49 +01:00
635ef78ec5
vulkan: work around Intel fp16 bug in mmq (#18814 )
2026-01-14 09:41:23 +01:00
7d587e5544
ggml-metal: do not copy headers for embedded, use current binary dir for embedded (#18705 )
2026-01-14 02:22:25 -05:00
d34aa07193
mmap: add Haiku support by skipping RLIMIT_MEMLOCK check (#18819 )
2026-01-14 08:11:05 +01:00
f709c7a33f
ci, tests : use cmake to download models and remove libcurl dependency (#18791 )
2026-01-14 07:46:27 +01:00
6e36299b47
llama : print_info alignment fix (#18708 )
2026-01-13 17:05:11 -06:00
60591f01d4
model : add EXAONE MoE (#18543 )
2026-01-14 07:28:38 +09:00
e4832e3ae4
vocab : fix attribute overrides for harmony (#18806 )
2026-01-13 17:40:13 +02:00
960e5e3b46
llama-mmap: fix direct-io loading fallback EOF exception (#18801 )
2026-01-13 15:57:07 +01:00
20ca2e12c4
model-conversion : remove -c 0 from model card template [no ci] (#18807 )
2026-01-13 14:13:10 +01:00
ea4a321f2a
HIP: add fattn-mma-f16 for RDNA4 (#18481 )
2026-01-13 20:52:16 +08:00
c1e79e610f
doc: ban AI-generated PR descriptions [no ci] (#18765 )
2026-01-13 13:43:12 +01:00
e047f9ee9d
mtmd: fix use_non_causal being reported incorrectly (#18793 )
2026-01-13 12:19:38 +01:00
0a57271ab6
CUDA : fix unused argument when USE_CUDA_GRAPH=OFF (#18800 )
2026-01-13 12:25:53 +02:00
076b0faf7d
graph : clean up t5 input builders (#18795 )
2026-01-13 01:43:51 -07:00
db79dc06b1
llama-bench: add direct_io parameter (#18778 )
2026-01-13 08:49:10 +01:00
537d4240d4
ci : remove libcurl in releases (#18775 )
2026-01-12 21:43:02 +01:00
bcf7546160
server : add arg for disabling prompt caching (#18776 )
2026-01-12 19:21:34 +02:00
36c5913c45
ci : use openssl for openEuler-latest-cmake-cann (#18779 )
2026-01-12 17:29:00 +01:00
8e649571cd
vendor : update cpp-httplib to 0.30.1 (#18771 )
2026-01-12 15:58:52 +01:00
4150da9a95
examples : add --kv-unified to batched example (#18774 )
2026-01-12 13:47:58 +01:00
8e2da778da
vulkan: change memory_logger to be controlled by an env var (#18769 )
2026-01-12 06:32:55 -06:00
ce3bf9b1a4
server: update docs for sleeping [no ci] (#18777 )
2026-01-12 13:01:24 +01:00
2bbe4c2cf8
vulkan: Use VK_EXT_shader_64bit_indexing to handle large mat_mul(_id) (#18678 )
2026-01-12 05:32:13 -06:00
1051ecd289
vulkan: Disable large coopmat matmul configuration on proprietary AMD driver (#18763 )
2026-01-12 07:29:35 +01:00
0c3b7a9efe
model: fix qwen3next broken due to #18683 (#18762 )
2026-01-11 21:00:10 +01:00
0e76501e1d
Vulkan: Optimize Matmul parameters for AMD GPUs with Coopmat support (#18749 )
2026-01-11 17:33:33 +01:00
4b060bf240
security: make it clear about subtopics in server (#18754 )
2026-01-11 16:51:03 +01:00
Xuan-Son Nguyen
손희준
Aman Gupta
shaofeiqi
Aleksei Nikiforov
Pádraic Slattery
Mariusz Woloszyn
Piotr Wilkin (ilintar)
Hendrik Erz
Daniel Bevenius
Jeff Bolz
Adrien Gallouët
Masato Nakasaka
Tarek Dakhran
Matthieu Coudron
Oliver Simons
Georgi Gerganov
ddh0
Sigbjørn Skjæret
Lennart Austenfeld
Julius Tischbein
Francisco Herrera
lhez
Reese Levine
Thore Koritzius
hipudding
Perry Naseck
Tarek Dakhran
Raul Torres
Chenguang Li
shaofeiqi
Johannes Gäßler
shalinib-ibm
Max Krasnyansky
Junwon Hwang
Ruben Ortlam
Daniel Benjaminsson
Ruben Ortlam
yulo
Gabe Goodhart
Radoslav Gerganov