This website requires JavaScript.
32e806b9c1
ci : fix check-release message parsing (#24751 )
2026-06-18 09:32:56 +02:00
6f1034b32a
[SYCL] support OPs: conv_2d, conv_2d_dw, conv2d_transpose (#24600 )
2026-06-18 14:40:03 +08:00
0b73fc79fe
ui: Update code formatting command in pre-commit hook (#24685 )
2026-06-18 08:33:50 +02:00
4a79037b8b
ci : fix Windows x64 (OpenVINO) release link (#24731 )
2026-06-17 23:30:08 -07:00
cae0a3b0b0
metal : check for BF16 support in concat kernel (#24747 )
2026-06-18 09:16:06 +03:00
f3e1828164
mtmd: llava_uhd should no longer use batch dim (#24732 )
2026-06-17 22:40:50 +02:00
2e88c49c90
ggml-cpu: Conditionally enable power11 backend based on compiler support (#24687 )
2026-06-18 00:15:19 +05:30
0843245cb1
metal : implement rope_back operator (#24725 )
2026-06-17 20:36:05 +03:00
8d2e580632
metal : add f16 and bf16 support for concat operator (#24724 )
2026-06-17 19:38:55 +03:00
4b4d13ae72
server: (router) add model management API (#23976 )
2026-06-17 18:04:58 +02:00
b4024af6c2
llama : skip main_gpu validation when no devices are available (#23405 )
2026-06-17 17:30:26 +03:00
1a2dea29b9
spec: fix segfault error on long prompts for eagle3 (#24707 )
2026-06-17 16:29:49 +02:00
74a80dd9c0
[SYCL] add dev2dev memcpy by SYCL API (#24476 )
2026-06-17 22:21:34 +08:00
d1759e4156
[SYCL] Add conv_3d (#24691 )
2026-06-17 22:20:01 +08:00
8086439a4c
webui: export conversations as jsonl (#24688 )
2026-06-17 13:25:47 +02:00
558e221b70
vulkan: record actual memory properties during buffer creation (#24326 )
2026-06-17 17:14:48 +08:00
ea21e03955
Revert "cuda: reset cuda context after reading memory size (#23935 )" (#24715 )
2026-06-17 10:59:35 +02:00
d5376cf5d7
ci: fix vulkan docker images (#24595 )
2026-06-17 09:43:45 +02:00
bae36efa30
UI : fix SSE transport detection and routing through CORS proxy. Assi… (#24500 )
2026-06-17 13:26:30 +07:00
51571722aa
opencl: optimize mul_mat_f16_f32_l4 for decode (#24504 )
2026-06-16 23:21:26 -07:00
cda63856b8
common: update logging to enforce max_capacity and optimize queue resizing (#24490 )
2026-06-16 23:19:11 -07:00
890f1a27ed
openvino: OV 2026.2, context-shift, Q5_1 support, gemma4 dense/embedding, and -fa off (#24503 )
2026-06-17 14:11:21 +08:00
58728bdbf0
sycl : Enable to support fp16 by OPs: SQR, SQRT, LOG, SIN, COS, CLAMP (#24692 )
2026-06-17 13:58:03 +08:00
ebbc1e51c1
SYCL: fix use-after-free bug with async memcpy in MoE prefill (#24676 )
2026-06-17 14:57:29 +09:00
9b260fc9ef
sycl: Add optional USM system allocations (#22526 )
2026-06-17 07:54:21 +02:00
74ade52741
vendor : update BoringSSL to 0.20260616.0 (#24693 )
2026-06-16 15:24:28 -03:00
c1304d7b28
ui: add source toggle to mermaid and svg blocks (#24652 )
2026-06-16 14:14:22 +02:00
02810c7aa8
Fix and restrict NVFP4 edge-cases in llama-graph (#24331 )
2026-06-16 11:52:38 +02:00
a1824902b5
spec: add backend sampling support for eagle3 (#24655 )
2026-06-16 11:05:52 +02:00
32120c10e3
vulkan: prefer host-visible memory buffers on UMA devices (#22930 )
2026-06-16 15:36:52 +08:00
d5fb104293
vulkan: Support gated_delta_net with S_v=16 (#24581 )
2026-06-16 02:26:57 -05:00
635b65ad7a
spec: add spec metrics mean acceptance length and acceptance rate per position (#24536 )
2026-06-16 09:23:09 +02:00
e3a74b2990
bench : add --offline (#24511 )
2026-06-16 08:26:05 +02:00
ac79caa7ce
sycl: support reordered Q4_K/Q5_K/Q6_K MoE MUL_MAT_ID (#24452 )
2026-06-16 00:35:00 -05:00
fdd109883d
[SYCL] Support OP EXPM1, support all UT cases of FLOOR, TRUNC, ROUND (#24363 )
2026-06-16 13:34:29 +08:00
4196b477da
sycl : Make GGML_SYCL_F16=ON the default (#23996 )
2026-06-15 22:34:02 -07:00
ad39ccaa19
vulkan: add col2im_1d op (#24425 )
2026-06-16 06:34:43 +02:00
7dad2f1a17
chat : fix LFM2 tool-call parsing double-escaping (#24667 )
2026-06-15 22:10:09 +02:00
e36a602ba3
mtmd: fix miscounting n_tokens (#24656 )
2026-06-15 18:07:14 +02:00
38d546330a
chat: include full unparsed prompt in debug (#24650 )
2026-06-15 17:33:54 +02:00
a1eb756c0b
docs: Add instructions to install llama.cpp from conda-forge (#22219 )
2026-06-15 17:12:25 +02:00
581e8eca8b
chat: harden peg-native tool call parsing (#24329 )
2026-06-15 15:37:04 +02:00
0ae3f450f0
chat: fix an "oldie but goodie" grammar generator bug that surfaced during last changes (#24653 )
2026-06-15 15:27:47 +02:00
e3cab403bf
mtmd : add post-decode callback (#24645 )
2026-06-15 16:02:05 +03:00
9dbc6621ae
vulkan: support more CONCAT types (#24579 )
2026-06-15 06:19:21 -05:00
6eab47181c
wasm : fix fallback symbol collision (#24639 )
2026-06-15 00:11:59 -07:00
e3bb1add8c
SYCL: use native subgroup size for K-quant DMMV (#21700 )
2026-06-15 03:10:53 -04:00
d8a3f523c8
sycl: fix soft_max_f32 max reduction (#24451 )
2026-06-15 15:10:12 +08:00
72be44f1d2
sycl : fix reorder function; add fp32/fp16 in build script (#24578 )
2026-06-15 15:08:34 +08:00
8872ab5467
sycl : enhance set_rows to support q1_0, mxfp4, nvfp4 (#24564 )
2026-06-15 15:01:40 +08:00
987fbd821d
[SYCL] add to support pool_1d, move pool_1d/2d code to pool.cpp/hpp (#24584 )
2026-06-15 15:01:07 +08:00
c035ff4902
[SYCL]: Remove per-allocation Level Zero runtime checks (#23399 )
2026-06-15 15:58:42 +09:00
272088b9f2
metal : add repeat bf16 (#24638 )
2026-06-15 09:57:16 +03:00
a6dff71270
chat: fix whitespace problems once and for all (#24624 )
2026-06-15 08:27:10 +02:00
2a6c391a5e
UI/svg block rendering (#24080 )
2026-06-15 08:11:36 +02:00
3686e9d643
CUDA: only support F32/F16 for GGML_OP_REPEAT (#24533 )
2026-06-15 14:11:00 +08:00
6e9007ae61
ggml-webgpu: improve i-quants mul_mat performance and speed up prefill (#24530 )
2026-06-15 10:15:30 +09:00
dd4623a74f
convert : fix lora base model arch retrieval (#24621 )
2026-06-15 00:55:26 +02:00
ef8268feee
fix(ui): render thinking/reasoning block content as markdown (#24611 )
2026-06-14 22:56:56 +02:00
5f04dc7ac3
ui: Add HEIC/HEIF image support (#24137 )
2026-06-14 12:42:16 -06:00
aedb2a5e9c
chat: add dedicated Cohere2MoE (North Code) parser (#24615 )
2026-06-14 20:17:40 +02:00
8edaca9034
docs : fix typos in CUDA-FEDORA.md and grammars/README.md (#24459 )
2026-06-14 23:03:38 +05:30
20c5266f8a
docker: specify registry to simplify Podman builds (#24607 )
2026-06-14 20:27:20 +03:00
fd5869fb62
UI/mobile keyboard and pwa popup fixes (#24610 )
2026-06-14 18:35:00 +02:00
1fd6dfe9f3
ui : fix ui clipping in mobile due to incorrect height setup (#24605 )
2026-06-14 22:15:51 +08:00
acd79d603c
jinja : add count/d/e filter aliases (#24606 )
2026-06-14 15:07:31 +02:00
6e14286eda
cli : fix not copying preserved tokens (#24258 )
2026-06-14 02:52:15 -07:00
8ed274ef46
Add cohere2moe to llama-vocab for TINY_AYA (#24601 )
2026-06-14 03:04:46 -04:00
46722116b9
ci : use CUDA label for cuda backend (#24594 )
2026-06-14 08:27:52 +02:00
c2ba3e47a2
add sycl to check-release (#24583 )
2026-06-14 03:42:26 +02:00
53bd47ea5b
ui : fix llama-ui-embed crash when no asset dir is given (#24597 )
2026-06-13 17:53:30 -05:00
4988f6e866
Add arch support for cohere2-MoE (#24260 )
2026-06-13 10:49:00 -07:00
f05cf4676a
jinja : fix negative step slice with start/stop values (#24580 )
2026-06-13 18:28:40 +02:00
e8067a8b36
ui: build-time gzip compression (#24571 )
2026-06-13 16:57:27 +02:00
341babcf73
jinja : fix split and replace with empty first arg (#24574 )
2026-06-13 16:56:59 +02:00
1a7718b4c5
vulkan: support non-contig unary/glu ops (#24215 )
2026-06-13 08:44:15 -05:00
597b6672e8
ui: keep original file name and path (#24568 )
2026-06-13 14:31:41 +02:00
57fe1f07c3
server: clean up static assets handling (#24550 )
2026-06-13 11:51:20 +02:00
d8a24ccee2
fit : wrap llama_device_memory_data (#24522 )
2026-06-13 08:09:52 +03:00
c34b92235b
fix sycl links in release notes (#24527 )
2026-06-13 03:37:55 +03:00
e37abd6b5f
mtmd: add batching API (#24384 )
2026-06-13 00:10:29 +02:00
f58bad4137
ci : unbreak release harder (#24545 )
2026-06-12 23:49:36 +02:00
cd5044661c
ci : unbreak release (#24544 )
2026-06-12 22:29:49 +02:00
ebc10770ac
server : fix reasoning budget WebUI precedence over model.ini (#24517 )
2026-06-12 17:59:56 +03:00
3e7bd4f39a
vulkan: add pipeline barriers for memcpy read operations (#23770 )
2026-06-12 16:43:50 +02:00
f7ca93d12c
ui: PWA support (#23871 )
2026-06-12 15:53:26 +02:00
02182fc5b9
fit : avoid including llama-ext.h in fit.h (#24506 )
2026-06-12 15:57:05 +03:00
f532be8fac
sync : ggml
2026-06-12 15:55:01 +03:00
e08c226a2c
ggml : bump version to 0.15.1 (ggml/1541)
2026-06-12 15:32:00 +03:00
70b54e140c
vendor : update cpp-httplib to 0.47.0 (#24395 )
2026-06-12 11:34:44 +02:00
6471e3c090
UI/jpeg exif orientation (#24196 )
2026-06-12 10:20:27 +02:00
88a39274ec
spec: add EAGLE3 speculative decoding support (#18039 )
2026-06-12 09:21:06 +02:00
85f99dca8b
ggml: support concat for scalar types at cuda backend (#24011 )
2026-06-12 14:32:44 +08:00
099ea76fb4
[SYCL] Fix CI build & release for SYCL backend (#24387 )
2026-06-12 14:30:24 +08:00
ba1df050f3
opencl: add q5_0/q5_1 gemm and gemv kernels for Adreno (#24319 )
2026-06-11 21:43:09 -07:00
1593d5684d
docker : support specifying the GCC version for CUDA (#24447 )
2026-06-12 05:12:09 +08:00
4c6595503f
vulkan: ifdef eMesaHoneykrisp (build fix) (#24479 )
2026-06-11 13:22:17 -05:00
263cc04a54
sync : ggml
2026-06-11 19:33:33 +03:00
17e59d6209
ggml : bump version to 0.15.0 (ggml/1539)
2026-06-11 19:32:38 +03:00
fdc3db9b65
vulkan: add fast path for contiguous buffer transfers (#23973 )
2026-06-11 21:46:25 +08:00
Sigbjørn Skjæret