| 2026-07-07 |
Georgi Gerganov | spec : fix naming, spacing (#25410) |
commit | commitdiff | tree |
| 2026-07-07 |
Oliver Simons | CUDA: Fuse MMVQ post-scale for NVFP4 (#24481) |
commit | commitdiff | tree |
| 2026-07-07 |
Alex | server : fix draft model fit vs load inconsistency... |
commit | commitdiff | tree |
| 2026-07-07 |
Thomas LECONTE | server : add timings and progress to /responses API... |
commit | commitdiff | tree |
| 2026-07-07 |
Thiago Padilha | server: enforce prompt cache RAM limit (#25070) |
commit | commitdiff | tree |
| 2026-07-07 |
zhangrunda | common : add missing <fstream> include in common.h... |
commit | commitdiff | tree |
| 2026-07-07 |
asf0 | ggml-hip : add -fno-finite-math-only alongside -ffast... |
commit | commitdiff | tree |
| 2026-07-07 |
Aman Gupta | llama: fix quantized kv-cache for dsv4 (#25202) |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | [SYCL] fix unsupported UT cases of CONT & CPY (#25231) |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | [SYCL] support op col2im_1d (#25264) |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | [SYCL] support OP cross_entropy_loss, cross_entropy_los... |
commit | commitdiff | tree |
| 2026-07-07 |
Todd Malsbary | sycl : set K_QUANTS_PER_ITERATION to 1 on DMMV path... |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | [SYCL] fix unsupport ACC UT cases for noncontiguous... |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | sycl : enhance argsort to support all UT cases (#25125) |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | sycl : use sycl func to fix AOT double type issue ... |
commit | commitdiff | tree |
| 2026-07-07 |
Neo Zhang | sycl : rename the env vars from "disable" to "enable... |
commit | commitdiff | tree |
| 2026-07-07 |
An Long | ggml : make ggml_time_init idempotent (#24422) |
commit | commitdiff | tree |
| 2026-07-07 |
o7si | speculative : fix out-of-bounds read in ngram-map on... |
commit | commitdiff | tree |
| 2026-07-07 |
fairydreaming | vulkan : check src0 type in GGML_OP_SET_ROWS to avoid... |
commit | commitdiff | tree |
| 2026-07-07 |
Hongqiang Wang | opencl: general flash attention decode performance... |
commit | commitdiff | tree |
| 2026-07-06 |
shalinib-ibm | common: Set optimal default thread count for ppc (... |
commit | commitdiff | tree |
| 2026-07-06 |
Pascal | metal: add col2im_1d op (f32/f16/bf16) (#25176) |
commit | commitdiff | tree |
| 2026-07-06 |
Johannes Gäßler | CUDA: remove -sm row, refactor cuBLAS (#24216) |
commit | commitdiff | tree |
| 2026-07-06 |
Pascal | server: fix deadlock in load_models() when erasing... |
commit | commitdiff | tree |
| 2026-07-06 |
Alexey Kopytko | CUDA: extend K-type validation to V-types for flash... |
commit | commitdiff | tree |
| 2026-07-06 |
Xuan-Son Nguyen | server: temporary skip model downloading API test ... |
commit | commitdiff | tree |
| 2026-07-06 |
ragz4125 | ggml-cpu: use UE4M3 LUT in ARM NVFP4 dot product (... |
commit | commitdiff | tree |
| 2026-07-06 |
shalinib-ibm | ggml-cpu: Enable tiled matmul on AIX (#25199) |
commit | commitdiff | tree |
| 2026-07-06 |
hokanosekai | vulkan: fix 32-bit integer overflow in CEIL_DIV (#25245) |
commit | commitdiff | tree |
| 2026-07-06 |
Pascal | ui: restore Ctrl+B sidebar toggle shortcut (#25307) |
commit | commitdiff | tree |
| 2026-07-06 |
Adrien Gallouët | scripts : use HF_TOKEN when downloading UI assets ... |
commit | commitdiff | tree |
| 2026-07-06 |
a-huk | ggml-hip: enable -ffast-math for HIP builds (#23862) |
commit | commitdiff | tree |
| 2026-07-06 |
Xuan-Son Nguyen | ui: fake 200 for proxy DELETE req (#25298) |
commit | commitdiff | tree |
| 2026-07-06 |
adavyas | ggml-cuda: optimize conv_transpose_1d indexing (#25310) |
commit | commitdiff | tree |
| 2026-07-05 |
Al G | Fix stale tensor-split params for draft models (#24814) |
commit | commitdiff | tree |
| 2026-07-05 |
Eve | abort if we see a multi buffer (#25276) |
commit | commitdiff | tree |
| 2026-07-05 |
liminfei-amd | ggml : fix tensor-parallel + -ncmoe crash on MoE models... |
commit | commitdiff | tree |
| 2026-07-05 |
Vexxie | ggml: Update VMM Pool allocation ggml-cuda.cu - Turing... |
commit | commitdiff | tree |
| 2026-07-05 |
fairydreaming | cuda : concat implementation for quantized types (... |
commit | commitdiff | tree |
| 2026-07-04 |
liminfei-amd | llama : add guard for K/V rotation input when buffer... |
commit | commitdiff | tree |
| 2026-07-04 |
Pascal | ui: add sync blocks so display/behavior settings can... |
commit | commitdiff | tree |
| 2026-07-04 |
fairydreaming | ggml : fix broken CPU concat implementation for quantiz... |
commit | commitdiff | tree |
| 2026-07-03 |
Piotr Wilkin... | chat: trim messages sent to StepFun parser (fixes long... |
commit | commitdiff | tree |
| 2026-07-03 |
Nick Towle | ui: Improve performance when streaming (#25225) |
commit | commitdiff | tree |
| 2026-07-03 |
Pascal | ui: strip path and weight extension from model id in... |
commit | commitdiff | tree |
| 2026-07-03 |
Ruixiang Wang | spec: support spec-draft-p-min in DFlash (#25246) |
commit | commitdiff | tree |
| 2026-07-03 |
Piotr Wilkin... | cuda: enable topk-moe fusion for 288 experts (#25267) |
commit | commitdiff | tree |
| 2026-07-03 |
Pascal | ui: align persisted config with strict server schema... |
commit | commitdiff | tree |
| 2026-07-03 |
Pascal | server + ui: ping silent SSE streams every 1s and kick... |
commit | commitdiff | tree |
| 2026-07-03 |
Aleksander... | ui: Add MCP Servers Opt-In for first time visitors... |
commit | commitdiff | tree |
| 2026-07-03 |
Gaurav Garg | Remove redundant CUDA copies after gated_delta_net... |
commit | commitdiff | tree |
| 2026-07-03 |
Alessandro... | vendor : update cpp-httplib to 0.49.0 (#25218) |
commit | commitdiff | tree |
| 2026-07-02 |
Adrien Gallouët | llama : add llama_model_ftype_name() (#25134) |
commit | commitdiff | tree |
| 2026-07-01 |
lhez | opencl: allow loading precompiled binary kernels from... |
commit | commitdiff | tree |
| 2026-07-01 |
Adrien Gallouët | common : use hf primary split as model path (#25194) |
commit | commitdiff | tree |
| 2026-07-01 |
Max Krasnyansky | hexagon: flash attention rework (optimizations, accurac... |
commit | commitdiff | tree |
| 2026-07-01 |
Johannes Gäßler | CUDA: consistent use of __restrict__ + PDL for FA ... |
commit | commitdiff | tree |
| 2026-07-01 |
ragz4125 | ggml-cpu: add AVX2 optimization for nvfp4 dot product... |
commit | commitdiff | tree |
| 2026-07-01 |
Aleksander... | ui Prevent tool messages from incorrectly appending... |
commit | commitdiff | tree |
| 2026-07-01 |
Aleksander... | ui: Remove PWA navigate fallback to prevent caching... |
commit | commitdiff | tree |
| 2026-07-01 |
lhez | opencl: initial q1_0 support (#25160) |
commit | commitdiff | tree |
| 2026-06-30 |
fairydreaming | cuda : prevent integer truncation and overflow errors... |
commit | commitdiff | tree |
| 2026-06-30 |
Jürgen Schmied | model : register t_layer_inp for qwen3next (#25141) |
commit | commitdiff | tree |
| 2026-06-30 |
Pascal | common,server: handle bracketed IPv6 literals in URL... |
commit | commitdiff | tree |
| 2026-06-30 |
Matt Jallo | CUDA: fix get_rows_back for tables with more than 65535... |
commit | commitdiff | tree |
| 2026-06-30 |
Johannes Gäßler | CUDA: fix Gemma E4B MTP FlashAttention (#25148) |
commit | commitdiff | tree |
| 2026-06-30 |
Kevin Liu | vulkan: roll bk loop in matmul for asahi linux (#24663) |
commit | commitdiff | tree |
| 2026-06-30 |
zduford | HIP: use hipBLAS for dense prefill on gfx900, keep... |
commit | commitdiff | tree |
| 2026-06-30 |
Masashi Yoshimura | ggml-webgpu: add support for NVFP4 (#25143) |
commit | commitdiff | tree |
| 2026-06-30 |
Oliver Simons | Revert "sched : reintroduce less synchronizations durin... |
commit | commitdiff | tree |
| 2026-06-29 |
Adrien Gallouët | common : dedup preset and cached model entries in ... |
commit | commitdiff | tree |
| 2026-06-29 |
Ruben Ortlam | vulkan: use flops instead of weight tensor size for... |
commit | commitdiff | tree |
| 2026-06-29 |
Aman Gupta | DeepSeek V4 (#24162) |
commit | commitdiff | tree |
| 2026-06-29 |
seryogakovalyov | tools/ui: restore Tailwind scanning in ignored worktree... |
commit | commitdiff | tree |
| 2026-06-29 |
o7si | common : remove unused regex-partial (#25118) |
commit | commitdiff | tree |
| 2026-06-28 |
Xuan-Son Nguyen | jinja, chat: add --reasoning-preserve flag (#25105) |
commit | commitdiff | tree |
| 2026-06-28 |
Aleksander... | Revert "ui: fix accessibility for hover-gated interacti... |
commit | commitdiff | tree |
| 2026-06-28 |
Pascal | ui: fix stop and reasoning skip in single-model mode... |
commit | commitdiff | tree |
| 2026-06-28 |
Ruixiang Wang | dflash: refactor draft model conversion (#25110) |
commit | commitdiff | tree |
| 2026-06-28 |
Aldehir Rojas | chat : implement minicpm5 parser (#24889) |
commit | commitdiff | tree |
| 2026-06-28 |
Xuan-Son Nguyen | jinja: add --dump-prog for debugging (#25086) |
commit | commitdiff | tree |
| 2026-06-28 |
Ruixiang Wang | spec : add DFlash support (#22105) |
commit | commitdiff | tree |
| 2026-06-28 |
Adrien Gallouët | common : allow --offline in llama download (#25091) |
commit | commitdiff | tree |
| 2026-06-28 |
Georgi Gerganov | logs : reduce v2 (#25078) |
commit | commitdiff | tree |
| 2026-06-27 |
Hongqiang Wang | opencl: flash attention improvement (#25069) |
commit | commitdiff | tree |
| 2026-06-27 |
Gaurav Garg | [CUDA] Added a cudaMemcpy2DAsync fast path to ggml_cuda... |
commit | commitdiff | tree |
| 2026-06-27 |
Neo Zhang | sycl : fix failed ut cases of norm (#25044) |
commit | commitdiff | tree |
| 2026-06-27 |
Ruben Ortlam | vulkan: fix step operator for 0 input (#25036) |
commit | commitdiff | tree |
| 2026-06-27 |
Christian Kastner | binaries : Improve rpc-server and export-graph-ops... |
commit | commitdiff | tree |
| 2026-06-27 |
Sigbjørn Skjæret | ci : add windows-openvino to check-release (#25022) |
commit | commitdiff | tree |
| 2026-06-27 |
Sigbjørn Skjæret | tests : fix test-chat-template --no-common option ... |
commit | commitdiff | tree |
| 2026-06-26 |
Adrien Gallouët | app : allow --version, --licenses & --help (#25054) |
commit | commitdiff | tree |
| 2026-06-26 |
Andreas Kieslinger | sched : reintroduce less synchronizations during split... |
commit | commitdiff | tree |
| 2026-06-26 |
Adrien Gallouët | devops : add llama in all docker images (#25035) |
commit | commitdiff | tree |
| 2026-06-26 |
Xuan-Son Nguyen | arg: fix handling --spec-draft-hf and --hf-repo-v ... |
commit | commitdiff | tree |
| 2026-06-26 |
Ravi Panchumarthy | openvino: Update to OV 2026.2.1, self-contained release... |
commit | commitdiff | tree |
| 2026-06-26 |
Georgi Gerganov | sync : ggml |
commit | commitdiff | tree |
| 2026-06-26 |
Georgi Gerganov | ggml : bump version to 0.15.3 (ggml/1550) |
commit | commitdiff | tree |
| 2026-06-26 |
nullname | vulkan: opt mul_mat_vecq for mi50 (#22933) |
commit | commitdiff | tree |
| 2026-06-26 |
Jiang, Fish | vulkan: add INTEL_XE1 arch enum and enable coopmat1... |
commit | commitdiff | tree |
| next |
overview / pkg / ggml / sources / llama.cpp / shortlog