wylab/llama.cpp
1
0
Fork 1
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-16 17:35:58 +02:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 7ea15bb64c vulkan: Improve build time for MSVC (#16545) b6761 Jeff Bolz 2025-10-14 07:51:36 -05:00
  • 9c7185dd28 CUDA: enable FA for FP32 KV cache (#16546) b6760 Johannes Gäßler 2025-10-14 14:22:47 +02:00
  • 1ee9d0b415 CUDA: use fastdiv + ggml_cuda_mad for mmvf (#16557) b6759 Aman Gupta 2025-10-14 19:16:21 +08:00
  • 48e2fa9fb7 CUDA: add fp kernel for larger batch size MoE (#16512) b6758 Aman Gupta 2025-10-14 19:15:15 +08:00
  • 5b6913c47b cuda : remove legacy copy-op pointer indirection code (#16485) b6757 Anav Prasad 2025-10-14 09:53:49 +00:00
  • bc07349a7f server : dynamic token limit for prompt cache (#16560) b6756 Georgi Gerganov 2025-10-14 08:48:50 +03:00
  • e60f241eac metal : FA support F32 K and V and head size = 32 (#16531) b6755 Georgi Gerganov 2025-10-13 23:07:57 +03:00
  • e38b7c6e9e graph : support cacheless embeddings with FA and iSWA (#16528) b6754 Georgi Gerganov 2025-10-13 22:42:37 +03:00
  • 5016b72862 opencl: fix build targeting CL 2 (#16554) b6753 lhez 2025-10-13 11:50:37 -07:00
  • 7049736b2d CUDA: fix numerical issues in tile FA kernel (#16540) b6752 Johannes Gäßler 2025-10-13 16:29:45 +02:00
  • 01d2bdc2bc ggml : fix build broken with -march=armv9-a on MacOS (#16520) b6751 Jie Fu (傅杰) 2025-10-13 20:48:47 +08:00
  • 56fc38b965 CANN: fix CPU memory leak in CANN backend (#16549) b6750 Chenguang Li 2025-10-13 17:01:24 +08:00
  • 1fb9504eb7 fix: add remark plugin to render raw HTML as literal text (#16505) Pascal 2025-10-13 10:55:32 +02:00
  • 3f750f8d76 metal: add support for opt_step_sgd (#16539) b6748 Sam/Samuel 2025-10-13 16:25:02 +08:00
  • c515fc5771 ggml : fix scalar path for computing norm (#16558) b6747 Georgi Gerganov 2025-10-13 11:22:27 +03:00
  • f9bc66c3eb CANN: Update several operators to support FP16 data format (#16251) b6746 hipudding 2025-10-13 08:52:22 +08:00
  • a31cf36ad9 metal : add opt_step_adamw and op_sum (#16529) b6745 Sam/Samuel 2025-10-13 02:43:14 +08:00
  • 81d54bbfd5 webui: remove client-side context pre-check and rely on backend for limits (#16506) Pascal 2025-10-12 18:06:41 +02:00
  • c7be9febcb [SYCL] fix UT fault cases: count-equal, argsort, pad OPs (#16521) b6743 Neo Zhang Jianyu 2025-10-12 21:53:35 +08:00
  • 8415f61e23 ci : add Vulkan on Ubuntu with default packages build (#16532) Mathieu Baudier 2025-10-12 15:48:03 +02:00
  • 2c301e91ab common : handle unicode during partial json parsing (#16526) b6741 Aldehir Rojas 2025-10-12 08:18:47 -05:00
  • 4b2dae383d common : update presets (#16504) Georgi Gerganov 2025-10-12 09:29:13 +03:00
  • 41aac5c69b ggml : Fix FP16 ELU positive branch (#16519) b6739 sirus20x6 2025-10-12 00:25:37 -05:00
  • a2fba89a42 hparams : add check for layer index in is_recurrent (#16511) b6738 Daniel Bevenius 2025-10-12 07:19:06 +02:00
  • 20cc625edc ggml: Correct SVE implementation in ggml_vec_dot_f16_unroll (#16518) b6737 sirus20x6 2025-10-12 00:15:00 -05:00
  • 11f0af5504 CUDA: faster tile FA, add oob checks, more HSs (#16492) b6736 Johannes Gäßler 2025-10-11 20:54:32 +02:00
  • a3cb04744f metal : fix mul-mm condition + fix mul-mv permuted kernels (#16494) b6735 Georgi Gerganov 2025-10-11 16:54:10 +03:00
  • 4a8fbe0a5e feat: render user content as markdown option (#16358) Pascal 2025-10-11 15:50:49 +02:00
  • 31d0ff1869 server / ranking : add sorting and management of top_n (#16403) b6733 Yann Follet 2025-10-11 21:39:04 +08:00
  • 97870e6497 cuda : avoid initializing unused devices (#16510) b6732 Diego Devesa 2025-10-11 04:02:26 -07:00
  • 477a66b035 convert : correctly handle LLaMA tokenizer for Jamba (#16470) amirai21 2025-10-11 11:33:41 +03:00
  • e60f01d941 server : fix division by zero when reporting stats (#16501) b6730 Georgi Gerganov 2025-10-10 22:15:05 +03:00
  • 81086cd6a3 vocab : mark EOT token for Granite models (#16499) b6729 Georgi Gerganov 2025-10-10 17:17:31 +03:00
  • 68ee98ae18 server : return HTTP 400 if prompt exceeds context length (#16486) b6728 Radoslav Gerganov 2025-10-10 17:11:07 +03:00
  • cdb6da468c server : log requests to /v1/completions (#16495) b6727 Radoslav Gerganov 2025-10-10 13:22:27 +03:00
  • 6d69ab3f26 cmake : Dont define XOPENSOURCE on AIX (#16481) b6726 Prajwal B Mehendarkar 2025-10-10 13:45:46 +05:30
  • 1faa13a118 webui: updated the chat service to only include max_tokens in the req… (#16489) b6725 Pascal 2025-10-09 22:54:57 +02:00
  • 1deee0f8d4 cpu : optimize the ggml NORM operation (#15953) b6724 duduta 2025-10-09 22:11:15 +03:00
  • d00cbea63c server : host-memory prompt caching (#16391) Georgi Gerganov 2025-10-09 18:54:51 +03:00
  • 8328fd4bae No markdown in cot (#16483) Pascal 2025-10-09 17:36:29 +02:00
  • 56b4795842 model-conversion : add support for SentenceTransformers (#16387) b6721 Daniel Bevenius 2025-10-09 14:35:22 +02:00
  • 2c0d875ae6 ci: add ARM64 Kleidiai build and test support (#16462) sudhiarm 2025-10-09 09:13:18 +01:00
  • aa4711d369 CANN: Improve ACL graph matching (#16166) b6719 Chenguang Li 2025-10-09 15:50:25 +08:00
  • d80d6d2400 kleidiai: kernel interface refactoring (#16460) b6718 Charles Xu 2025-10-09 09:29:17 +02:00
  • b260213755 [SYCL] refactor soft_max, add soft_max_back (#16472) b6717 Neo Zhang Jianyu 2025-10-09 15:25:11 +08:00
  • e08db42595 model: EmbeddingGemma Adding Support for SentenceTransformers Dense Modules (#16367) Saba Fallah 2025-10-09 08:39:18 +02:00
  • 12bbc3fa50 refactor: centralize CoT parsing in backend for streaming mode (#16394) b6715 Pascal 2025-10-08 22:18:41 +02:00
  • 9d0882840e Disable CUDA host buffers on integrated GPUs (#16308) b6714 ai-fonsi 2025-10-08 20:21:46 +02:00
  • d2ee056e1d server : fix cancel pending task (#16467) b6713 issixx 2025-10-08 17:20:18 +09:00
  • b2c08c9ec4 metal : mark FA blocks (#16372) Georgi Gerganov 2025-10-08 10:57:53 +03:00
  • 7fdd16b432 server : improve context checkpoint logic (#16440) b6711 Georgi Gerganov 2025-10-08 10:57:29 +03:00
  • 74b8fc17f9 ggml webgpu: profiling, CI updates, reworking of command submission (#16452) b6710 Reese Levine 2025-10-07 13:48:56 -07:00
  • aeaf8a36f0 llama : support LiquidAI LFM2-MoE hybrid model (#16464) b6709 Tarek Dakhran 2025-10-07 20:03:35 +02:00
  • df1b612e29 server : add /v1/health endpoint (#16461) b6708 Georgi Gerganov 2025-10-07 15:57:14 +03:00
  • 4e0388aa8a webui : added download action (#13552) (#16282) Sascha Rogmann 2025-10-07 11:11:08 +02:00
  • ef4c5b87ea presets : fix pooling param for embedding models (#16455) b6706 Georgi Gerganov 2025-10-07 10:32:32 +03:00
  • c61ae20d05 rpc : update documentation (#16441) Radoslav Gerganov 2025-10-07 09:59:13 +03:00
  • 0123ff38f5 memory : use sequential equal splits for recurrent modules (#16442) b6704 Georgi Gerganov 2025-10-07 08:24:17 +03:00
  • 0a319bb75e metal : add support for non-padded FA KV (#16148) b6703 Georgi Gerganov 2025-10-07 08:23:30 +03:00
  • 1d6092fc72 tests : add -INF blocks to the KQ mask in the FA tests (#16380) b6702 Georgi Gerganov 2025-10-07 08:22:35 +03:00
  • 8ae32dc9ec metal : various optimizations + refactoring (#16446) b6701 Georgi Gerganov 2025-10-07 08:21:40 +03:00
  • 3df2244df4 llama : add --no-host to disable host buffers (#16310) b6700 Gadflyii 2025-10-06 12:55:53 -05:00
  • c08002a198 chat : Granite Docling stopping (#16438) b6699 Gabe Goodhart 2025-10-06 10:59:40 -06:00
  • 3a002afafa ci : refactor sdk caching to minimize storage (#16414) Sigbjørn Skjæret 2025-10-06 17:40:21 +02:00
  • a23b9bdbd3 ggml : fix unaligned access in AMX code (#16315) b6697 Georgi Gerganov 2025-10-06 16:05:27 +03:00
  • 04e632a4aa ci : remove missing reranker model files (#16444) Daniel Bevenius 2025-10-06 14:56:59 +02:00
  • a80ff183ab ggml-cpu : fix leftover handling in ggml_vec_scale_f32 for SVE (#16443) b6695 Daniel Bevenius 2025-10-06 14:17:12 +02:00
  • 1d49ca3759 nix : removed metal for nix (#16118) Yuannan 2025-10-06 09:29:56 +00:00
  • c5fef0fcea server: update readme to mention n_past_max metric (#16436) Oleksandr Kuvshynov 2025-10-06 03:53:31 -04:00
  • ca71fb9b36 model : Granite docling + Idefics3 preprocessing (SmolVLM) (#16206) b6692 Gabe Goodhart 2025-10-05 06:57:47 -06:00
  • 35266573b9 ggml webgpu: actually add softmax, fix rms_norm offset (#16400) b6691 Reese Levine 2025-10-04 20:59:31 -07:00
  • 86df2c9ae4 vulkan: use a more appropriate amount of threads when generating shaders (#16418) b6690 Eve 2025-10-04 20:04:27 +00:00
  • f39283960b rpc : check src buffer when copying tensor (#16421) b6689 Radoslav Gerganov 2025-10-04 16:22:45 +03:00
  • 898acba681 rpc : add support for multiple devices (#16276) b6688 Radoslav Gerganov 2025-10-04 12:49:16 +03:00
  • e29acf74fe vulkan : incremental shader builds (#16341) b6687 Acly 2025-10-04 11:42:56 +02:00
  • 128d522c04 chat : support Magistral thinking (#16413) b6686 Pascal 2025-10-03 20:51:48 +02:00
  • f6dcda3900 server : context checkpointing for hybrid and recurrent models (#16382) b6685 ddh0 2025-10-03 13:34:51 -05:00
  • 606a73f531 metal : fix loop bound in ggml_mem_ranges (#16412) b6684 Georgi Gerganov 2025-10-03 19:18:56 +03:00
  • 946f71ed9a llama : fix shapes for bert/mpt q/k norm (#16409) b6683 Sigbjørn Skjæret 2025-10-03 14:40:25 +02:00
  • 638d330246 ggml : fix graph reallocation with multiple chunks (#16396) b6682 Acly 2025-10-03 13:49:08 +02:00
  • 84c8e305e8 Fix missing messages on sibling navigation (#16408) Aleksander Grygier 2025-10-03 12:51:40 +02:00
  • 2aaf0a2a20 vulkan: Replace uses of maxMemoryAllocationSize and VK_WHOLE_SIZE (#16354) b6680 Jeff Bolz 2025-10-03 05:50:46 -05:00
  • 0e1f838556 vulkan: Fix FA coopmat1 invalid array indexing (#16365) b6679 Jeff Bolz 2025-10-03 04:52:46 -05:00
  • ad126479c2 ci : change macos-13 to macos-15-intel (#16401) b6678 Daniel Bevenius 2025-10-03 11:45:16 +02:00
  • 77233277c9 Capture model name only after first token (streaming) or completed request (#16405) Aleksander Grygier 2025-10-03 11:30:39 +02:00
  • e308efda8e vulkan: in flash attention, bounds check against nem1 (don't rely on GGML_KQ_MASK_PAD) (#16316) b6676 Jeff Bolz 2025-10-03 03:33:08 -05:00
  • 136bda78c5 webui : Fix messages payload sent to chat completions (#16402) Aleksander Grygier 2025-10-03 09:11:34 +02:00
  • 5113efd34c fix: track viewportHeight via window.innerHeight to avoid unwanted scrolling (#16356) Pascal 2025-10-03 08:01:31 +02:00
  • d64c8104f0 test-barrier : do not use more threads than physically available (#16389) b6673 Sigbjørn Skjæret 2025-10-02 20:10:12 +02:00
  • ef07a40906 ggml webgpu: add support for soft_max, optimize rms_norm (#16357) b6672 Reese Levine 2025-10-02 11:00:31 -07:00
  • 34fcc5a4ac model : Apertus model implementation (#15852) b6671 Piotr Wilkin (ilintar) 2025-10-02 19:43:22 +02:00
  • 91a2a56556 musa: update compile flags (#16265) b6670 R0CKSTAR 2025-10-02 21:29:56 +08:00
  • 72ee736c44 ci : fix ubuntu-latest-cmake-rpc (disable ccache) (#16388) Sigbjørn Skjæret 2025-10-02 13:51:36 +02:00
  • f09aefaa84 ci: update vulkan ci (#16294) Eve 2025-10-02 08:10:07 +00:00
  • bbd32bc038 ci : fix clean-up of old logs (#16381) Georgi Gerganov 2025-10-02 10:35:43 +03:00
  • 2be72c2b12 SYCL: Update to oneAPI 2025.2 (#16371) b6666 Neo Zhang Jianyu 2025-10-02 15:16:25 +08:00
  • 95ce098544 HIP: add IMbackK to codeowner (#16375) b6665 uvos 2025-10-02 05:52:59 +02:00
  • c8dedc9999 CI: reenable cdna in rocm docker builds (#16376) uvos 2025-10-01 23:32:39 +02:00
  • e95fec640f HIP: Disable ROCWMMA fattn on CDNA when compiled against ROCWMMA 2.0.0 (#16221) b6663 uvos 2025-10-01 23:09:25 +02:00
  • ded67b9444 llama : parameter conversion and loading fixes for PLaMo2 variants (#16075) b6662 Shunta Saito 2025-10-02 06:08:15 +09:00