-
03914c7ef8
common : move all common_chat_parse_* to chat-parser.cpp. (#17481)
DAN™
2025-11-28 13:29:36 -05:00
-
3ce7a65c2f
server: fix: /metrics endpoint returning JSON-escaped Prometheus format (#17386)
o7si
2025-11-29 02:14:00 +08:00
-
e072b2052e
ggml : add GGML_SCHED_NO_REALLOC option to disable reallocations in ggml_backend_sched (#17276)
Diego Devesa
2025-11-28 07:33:23 -08:00
-
c6f7a423c8
[MUSA] enable fp16/fast_fp16/bf16_mma on PH1 (#17551)
R0CKSTAR
2025-11-28 21:08:29 +08:00
-
2e7ef98f18
ggml-cuda: add stricter checking for fusion (#17568)
Aman Gupta
2025-11-28 20:34:51 +08:00
-
ddf9f94389
server : add Anthropic Messages API support (#17570)
Fredrik Hultin
2025-11-28 12:57:04 +01:00
-
ff55414c42
model : Qwen3 Next (#16095)
Piotr Wilkin (ilintar)
2025-11-28 12:02:56 +01:00
-
73955f7d2a
CUDA: no FP16 arithmetic for vector FA kernel (#17558)
Johannes Gäßler
2025-11-28 10:29:09 +01:00
-
35cf8887e1
vulkan: Implement GGML_OP_TRI (#17503)
Jeff Bolz
2025-11-28 03:07:29 -06:00
-
15d2b46b4d
rpc : cache and reuse compute graphs (#15405)
Radoslav Gerganov
2025-11-28 10:33:51 +02:00
-
6bca76ff5e
HIP: enable mul_mat_f for RDNA4 (#17437)
yulo
2025-11-28 15:24:30 +08:00
-
cd0e3a7a3b
SOLVE_TRI CUDA kernel for small matrices (#17457)
Piotr Wilkin (ilintar)
2025-11-28 05:15:32 +01:00
-
efaaccdd69
refactor pad_reflect_1d to make the UT case pass (#17204)
Neo Zhang Jianyu
2025-11-28 08:50:56 +08:00
-
4abef75f2c
vulkan: Implement SOLVE_TRI (#17486)
Jeff Bolz
2025-11-27 08:48:00 -06:00
-
c386114922
arch : add description about LLM_TENSOR_INFOS (#17550)
Georgi Gerganov
2025-11-27 16:34:13 +02:00
-
6783b11fb0
models : fix LFM2 tensors (#17548)
Georgi Gerganov
2025-11-27 16:04:29 +02:00
-
909072abcf
cuda : fix UMA detection on discrete GPUs. (#17537)
matt23654
2025-11-27 11:35:35 +00:00
-
cd8370b408
ggml-cpu: aarm64: q4_K repack gemm and gemv implementations (dotprod only) (#17494)
Alberto Cabrera Pérez
2025-11-27 11:25:14 +00:00
-
d21a76ac38
devops: Add build-essential to Ubuntu 26.04 image (#17531)
Eric Curtin
2025-11-27 10:35:47 +00:00
-
4fcd87cf7c
gguf-py : skip endian-conversion of MXFP4 data (#17523)
Aleksei Nikiforov
2025-11-27 11:35:38 +01:00
-
b78db3bd50
vulkan : move contiguous checks to device_supports_op (#17490)
Acly
2025-11-27 06:54:19 +01:00
-
142df17c9c
vulkan: use a fixed 1KB buffer for the add_rms_fusion opt (#17514)
Jeff Bolz
2025-11-26 23:32:30 -06:00
-
e509411cf1
server: enable jinja by default, update docs (#17524)
Xuan-Son Nguyen
2025-11-27 01:02:50 +01:00
-
7cba58bbea
opencl: add sqr, sqrt, mean and ssm_conv (#17476)
lhez
2025-11-26 13:29:58 -08:00
-
5449367b21
Fix chunks being too small with small matrix sizes (#17526)
Alberto Cabrera Pérez
2025-11-26 21:14:54 +00:00
-
1d594c295c
clip: (minicpmv) fix resampler kq_scale (#17516)
Han Qingzhe
2025-11-27 04:44:07 +08:00
-
eec1e33a9e
vulkan: allow graph_optimize for prompt processing workloads (#17475)
Jeff Bolz
2025-11-26 09:46:33 -06:00
-
879d673759
vulkan: Implement top-k (#17418)
Jeff Bolz
2025-11-26 09:45:43 -06:00
-
6ab4e50d9c
ggml-cpu : add RISC-V Zvfh impl for ggml_vec_mad_f16 (#17448)
xctan
2025-11-26 21:33:05 +08:00
-
2336cc4784
cmake : use EXCLUDE_FROM_ALL to avoid patch-boringssl.cmake (#17520)
Adrien Gallouët
2025-11-26 14:15:21 +01:00
-
e6923caaec
ggml : fix ARM feature verification (#17519)
Adrien Gallouët
2025-11-26 14:14:41 +01:00
-
3e18dba9fd
HIP: Patch failed testcase in WMMA-MMQ kernels for RDNA 4 (#17502)
Jiacheng (Jason) Chen
2025-11-26 05:18:48 -05:00
-
eeb5605de2
CANN: Add MROPE and IMROPE support (#17401)
hipudding
2025-11-26 16:44:19 +08:00
-
f3a848a3b1
chore: upgrade cpp-httplib from v0.27.0 to v0.28.0 (#17513)
o7si
2025-11-26 15:21:06 +08:00
-
b3b03a7baf
vulkan: Implement GGML_OP_CUMSUM (#17479)
Jeff Bolz
2025-11-26 00:08:10 -06:00
-
583cb83416
ggml : add ggml_top_k (#17365)
Georgi Gerganov
2025-11-25 15:31:43 +02:00
-
05872ac885
convert : fix big-endian conversion (#17431)
Aleksei Nikiforov
2025-11-25 14:18:16 +01:00
-
55ab25caf5
codeowners : remove slaren (#17492)
Diego Devesa
2025-11-25 04:00:23 -08:00
-
064c90d843
CANN: supports out_prod operator for F32 and F16 (#17406)
TianHao324
2025-11-25 17:39:06 +08:00
-
b1846f1c8e
webui: add rehype plugin to restore HTML in Markdown table cells (#17477)
Pascal
2025-11-25 08:01:02 +01:00
-
d414db02d3
vulkan: Use fewer rows for scalar FA when HS is not a multiple of 16 (#17455)
Jeff Bolz
2025-11-25 00:11:27 -06:00
-
877566d512
llama: introduce support for model-embedded sampling parameters (#17120)
Aaron Teo
2025-11-25 09:56:07 +08:00
-
3d07caa99b
vulkan: more FA details in vk_perf_logger (#17443)
Jeff Bolz
2025-11-24 15:25:24 -06:00
-
134e6940ca
llama : skip output reordering for single token batches (#17466)
Daniel Bevenius
2025-11-24 21:06:17 +01:00
-
0543f928a3
HIP: WMMA-MMQ kernels for RDNA 4 (#17156)
Jiacheng (Jason) Chen
2025-11-24 14:00:10 -05:00
-
b61de2b2df
convert : allow quantizing lora again (#17453)
Sigbjørn Skjæret
2025-11-24 15:50:55 +01:00
-
b8372eecd9
server: split server.cpp code into server/common/task/queue (#17362)
Xuan-Son Nguyen
2025-11-24 14:41:53 +01:00
-
6ab8eacddf
examples : add -kvu to batched usage example [no ci] (#17469)
Daniel Bevenius
2025-11-24 14:38:45 +01:00
-
2d50b9d8cb
sync : ggml
Georgi Gerganov
2025-11-24 14:28:37 +02:00
-
697edfeead
ggml : remove dirty flag from version string (ggml/1391)
Daniel Bevenius
2025-11-24 12:51:50 +01:00
-
dbb852b549
ggml-cpu: arm64: q4_K repack gemm and gemv implementations (i8mm) (#16739)
Alberto Cabrera Pérez
2025-11-24 11:08:11 +00:00
-
5f55c385cb
ggml: add RISC-V cpu-feats (#17461)
ixgbe
2025-11-24 19:07:14 +08:00
-
4902eebe33
models : Added support for RND1 Diffusion Language Model (#17433)
william pan
2025-11-23 22:16:56 -08:00
-
923ae3c619
hexagon: add support for ROPE_NEOX (#17458)
Max Krasnyansky
2025-11-23 18:55:56 -08:00
-
01ad35e6d6
CANN: Define
cann_graph_update_required before macro (#17434)
Raul Torres
2025-11-24 02:02:52 +00:00
-
fcb013847c
ggml-hexagon: Initial Hexagon v68/v69 support (#17394)
M. Mediouni
2025-11-24 01:54:49 +01:00
-
d5bc1ad110
ggml-hexagon: add
hex_supported_buffer for better buffer supported check (#17212)
nullname
2025-11-24 06:26:36 +08:00
-
0c7220db56
webui: minor settings reorganization and add disable autoscroll option (#17452)
Pascal
2025-11-23 18:42:00 +01:00
-
96ac5a2329
cuda : support non-contiguous i32 to i32 copy (#17326)
Sigbjørn Skjæret
2025-11-23 11:13:34 +01:00
-
bc809e9c53
vulkan: Update docker image to Ubuntu 26.04 to enable glslc features (#17439)
Eric Curtin
2025-11-23 09:29:36 +00:00
-
54d83bbe85
vulkan: remove a couple unnecessary switches (#17419)
Jeff Bolz
2025-11-22 23:29:40 -06:00
-
4949ac0f18
ci : switch to BoringSSL on Server workflow (#17441)
Adrien Gallouët
2025-11-22 21:38:19 +01:00
-
3f3a4fb9c3
Revive MUL_MAT_ID to perf testing (#17397)
Masato Nakasaka
2025-11-22 18:55:43 +09:00
-
028f93ef98
HIP: RDNA4 tensor core support for MMF (#17077)
yulo
2025-11-22 07:03:24 +08:00
-
8e9ddba610
opencl: refine condition for kqv mm (#17392)
lhez
2025-11-21 14:34:48 -08:00
-
23bc779a6e
model : detect GigaChat3-10-A1.8B as deepseek lite (#17420)
ubergarm
2025-11-21 08:51:38 -05:00
-
28175f857d
cmake : add option to build and link BoringSSL (#17205)
Adrien Gallouët
2025-11-21 11:46:45 +01:00
-
9cc4080441
ci : start using OpenSSL (#17235)
Adrien Gallouët
2025-11-21 11:45:00 +01:00
-
f1ffbba68e
vulkan: disable async for older Intel devices (#17369)
Jeff Bolz
2025-11-21 02:58:17 -06:00
-
2370665e56
CANN: Refactor
evaluate_and_capture_cann_graph (#17333)
Raul Torres
2025-11-21 08:23:29 +00:00
-
21d31e0810
ggml-hexagon: fix swiglu failure at
test-backend-ops (#17344)
nullname
2025-11-21 07:45:05 +08:00
-
dd0f321941
readme : add Unsloth exporting to GGUF in tools (#17411)
Daniel Han
2025-11-20 11:07:36 -08:00
-
054a45c3d3
grammar: fix regression caused by #17381 (#17412)
Xuan-Son Nguyen
2025-11-20 18:35:10 +01:00
-
4c91f2633f
Improved file naming & structure for UI components (#17405)
Aleksander Grygier
2025-11-20 14:07:31 +01:00
-
92c0b387a9
grammar : fix integer overflow (#17381)
Piotr Wilkin (ilintar)
2025-11-20 13:47:04 +01:00
-
2286a360ff
sync : ggml
Georgi Gerganov
2025-11-20 14:09:48 +02:00
-
1d321e592b
metal : fix compile on macos 11 (whisper/3533)
YangLe
2025-11-20 19:54:54 +08:00
-
196f5083ef
common : more accurate sampling timing (#17382)
Georgi Gerganov
2025-11-20 13:40:10 +02:00
-
5088b435d4
convert : fix TypeError when loading base model remotely in convert_lora_to_gguf (#17385)
o7si
2025-11-20 19:30:12 +08:00
-
845f200b28
ggml : Fix transposed SOLVE_TRI result (#17323)
Piotr Wilkin (ilintar)
2025-11-20 11:58:21 +01:00
-
a7784a8b1d
DGX Spark: UMA support (#17368)
Scott Fudally
2025-11-20 02:32:02 -08:00
-
79bb743512
ggml : remove useless and error-prone variadic macros (#17399)
Adrien Gallouët
2025-11-20 11:18:27 +01:00
-
3ae282a06f
kleidiai: fix zero-size array declaration (#17240)
sudhiarm
2025-11-20 09:45:49 +00:00
-
5be353ec4a
ggml-cpu:add RISC-V RVV (Zvfh) optimization for FP16 vector scaling (#17314)
ixgbe
2025-11-20 14:09:18 +08:00
-
7d77f07325
vulkan: implement ADD1, ARANGE, FILL, SOFTPLUS, STEP, ROUND, CEIL, FLOOR, TRUNC (#17319)
Giuseppe Scrivano
2025-11-19 17:29:45 +01:00
-
1fa4551af0
vulkan: support larger argsort (#17313)
Jeff Bolz
2025-11-19 10:25:50 -06:00
-
2eba631b81
vulkan: Add copy_transpose shader (#17371)
Jeff Bolz
2025-11-19 09:50:43 -06:00
-
99c53d6558
webui: Add a "Continue" Action for Assistant Message (#16971)
Aleksander Grygier
2025-11-19 14:39:50 +01:00
-
07b0e7a5ac
convert : use self.block_count everywhere instead of reading hparams (#17359)
Sigbjørn Skjæret
2025-11-19 11:52:38 +01:00
-
fd7353d5eb
cuda: fix rope fusion for gemma3 (#17378)
Aman Gupta
2025-11-19 18:25:05 +08:00
-
6fd4f95367
Fix too relaxed check on CUDA "fast copy" (can_be_transposed) condition (#17332)
Piotr Wilkin (ilintar)
2025-11-19 10:36:33 +01:00
-
980b7cd17e
vulkan: force full subgroups for flash attention to fix intel subgroup crash (#17356)
Ruben Ortlam
2025-11-19 08:46:26 +01:00
-
c49daff5ba
ggml-cpu: Don't pass -mpowerpc64 when -mcpu already implies it (#17308)
Jeremy Rand
2025-11-19 06:19:00 +00:00
-
10e9780154
chat: fix int overflow, prevent size calculation in float/double (#17357)
Xuan-Son Nguyen
2025-11-18 19:11:53 +01:00
-
a045492088
vocab : call reserve() for building plamo-2-translate suffix (#17343)
Haiyue Wang
2025-11-19 01:58:22 +08:00
-
1920345c3b
common : Generalized XML-style tool-call parsing with streaming support (GLM 4.5/4.6 + MiniMax M2 + SeedOSS + Kimi-K2 + Qwen3-Coder + Apriel-1.5 + Xiaomi-MiMo) (#16932)
hksdpc255
2025-11-19 04:54:15 +11:00
-
561a3e2788
ci : change the openEuler-310p image to fix release (#17361)
jiahao su
2025-11-19 01:10:23 +08:00
-
f40a2e5f11
gitignore : be more specific about ignored stuff (#17354)
Georgi Gerganov
2025-11-18 16:44:53 +02:00
-
bc4064cfea
CANN: fix acl_tensor_ptr usage in ASCEND_310P ROPE (#17347)
Chenguang Li
2025-11-18 16:41:52 +08:00
-
97cb3fd5ae
fix: resolve undefined variable 'svr' compilation error (#17348)
o7si
2025-11-18 16:10:47 +08:00