happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 925a83ac70 snapshot: debug ggml-hexagon swiglu-oai shouyud 2025-12-16 14:12:43 -0500
  • 1216d23f31 Missing generic fallbacks for x86 and powerpc Alberto Cabrera 2025-12-16 19:04:09 +0000
  • 0879d22196 Adding maybe unused keyword for Mac and Windows. JTischbein 2025-12-16 20:00:46 +0100
  • 1bc37a0ba7 server: (router) allow child process to report status via stdout Xuan Son Nguyen 2025-12-16 19:33:38 +0100
  • ef83fb8601
         model: fix LFM2 missing tensors (#18105) b7438 Xuan-Son Nguyen 2025-12-16 19:07:43 +0100
  • 40d79fde96
         Merge a70e4c87e4 into ec98e20021 0Marble 2025-12-16 12:24:00 -0500
  • 128a6c2831 ggml-cpu: add DELTA_NET backend + tests hauhaut 2025-12-16 18:21:29 +0100
  • 529c5b440c
         Merge 8b7a750411 into ec98e20021 Grigore Mihai 2025-12-16 19:11:51 +0200
  • ac5667dcc6 fix eagle3 logits sync bug & remove ggml_set_sync() ruixiangw 2025-12-16 16:53:28 +0000
  • 82c6ad3b24
         Merge 3a451fc845 into ec98e20021 Mishusha 2025-12-16 17:43:29 +0100
  • 72a41fd960 fix missing tensor tarek/feat/lfm2-asr-upstream Xuan Son Nguyen 2025-12-16 17:34:20 +0100
  • 653b544c0f model: fix LFM2 missing tensors Xuan Son Nguyen 2025-12-16 17:34:02 +0100
  • f6d79fe1b1 Remove branching in llama-model-loader.cpp and reduce code duplications in llama-mmap.cpp JTischbein 2025-12-16 13:58:43 +0100
  • 7865a1519e Merge branch 'master' into tarek/feat/lfm2-asr-upstream Xuan Son Nguyen 2025-12-16 17:30:59 +0100
  • 87e4a00c4c minor - added GLM-4.6V to big tests - added missing deps for python test Saba Fallah 2025-12-16 17:28:46 +0100
  • 90ec9d1bee improve tool calling outside of reasoning blocks, improve code interpreter documentation around async Josh Leverette 2025-12-15 20:30:51 -0600
  • f7f6040a78 smoother auto-scroll handling Josh Leverette 2025-12-15 09:58:43 -0600
  • ca11511dd4 Fix message stats for combined messages Josh Leverette 2025-12-15 09:43:33 -0600
  • 1bbf328caf enhance context retention between tool calls, fix lints Josh Leverette 2025-12-15 09:29:10 -0600
  • 0a428ff112 webui: Client-side implementation of tool calling with calculator tool and (javascript) code interpreter tool Josh Leverette 2025-12-14 18:24:18 -0600
  • a3ebc93d71 remove some redundant ggml_cont Xuan Son Nguyen 2025-12-16 17:20:13 +0100
  • 80edc59b57 llama-fit-params: force disable mlock Johannes Gäßler 2025-12-16 17:01:27 +0100
  • cea578bc8c rename functions to conformer Xuan Son Nguyen 2025-12-16 16:58:00 +0100
  • 00d235700d Merge remote-tracking branch 'sfallah/master' into sf/deepseek-ocr Saba Fallah 2025-12-16 16:45:43 +0100
  • 0a192937a1 qwen3next: trim comments hauhaut 2025-12-16 02:08:50 +0100
  • 4114537fb9 ggml-cuda: Delta-Net linear attention for Qwen3-Next hauhaut 2025-12-16 02:08:41 +0100
  • 92175d33a0 llama-fit-params: lower ctx size for multi GPU Johannes Gäßler 2025-12-16 16:36:57 +0100
  • 0548962e80
         Allow converting multi-tensor models from read-only locations Yuri Khrustalev 2025-12-16 10:01:55 -0500
  • 952877ec24 chore: reformat code with clang-formatter to pass cli test shouyud 2025-12-16 08:41:54 -0500
  • 05693357c8 refactor: cleanup commented unused code shouyud 2025-12-16 08:31:16 -0500
  • e51b6bf2b9 Revert "debug: temporarily disable unnecessary log message for debug purpose" shouyud 2025-12-16 08:27:57 -0500
  • 4fc63e1b63 llama-fit-params: fix underflow for dense models Johannes Gäßler 2025-12-16 14:27:30 +0100
  • cad07fa4b5 fix gramma and empty spaces zhang hui 2025-12-16 21:27:14 +0800
  • ec98e20021
         llama: fix early stop in params_fit if ctx is set (#18070) b7437 Johannes Gäßler 2025-12-16 14:24:00 +0100
  • 512b2c8fe4 merge with changes from https://github.com/ggml-org/llama.cpp/pull/18042 Saba Fallah 2025-12-16 14:07:04 +0100
  • dfa79a9484
         Merge branch 'master' into quantize Ed Addario 2025-12-16 13:57:54 +0100
  • c3b6685599
         Merge branch 'master' into imatrix Ed Addario 2025-12-16 13:57:27 +0100
  • 9311aa50a7
         Merge 8c252d13b8 into 59977eba7b DAN™ 2025-12-16 14:46:14 +0200
  • bbd234e2bf llama-fit-params: QoL impr. for prints/errors Johannes Gäßler 2025-12-16 10:59:15 +0100
  • 26ef6770d0 Cleanup, consistent style Alberto Cabrera 2025-12-16 12:33:26 +0000
  • 59977eba7b
         server: fix crash when batch > ubatch with embeddings (#17912) b7436 yifant-code 2025-12-16 07:27:36 -0500
  • ea0dec4348
         Update tools/server/server.cpp Georgi Gerganov 2025-12-16 14:27:20 +0200
  • 031b053cd3
         Merge branch 'master' into fix/embedding-batch-validation Georgi Gerganov 2025-12-16 14:26:40 +0200
  • 815123c472 Removed non-correct unused variables statements Alberto Cabrera 2025-12-16 12:25:10 +0000
  • 79dbae034a
         model-conversion : remove -fa option in model card template [no ci] (#18088) Daniel Bevenius 2025-12-16 13:25:09 +0100
  • 7f2b2f3c77
         arch: refactor LLM_TENSOR_NAMES (#18051) b7434 Xuan-Son Nguyen 2025-12-16 13:22:30 +0100
  • 3d20246c0d
         model-conversion : remove -fa option in model card template [no ci] Daniel Bevenius 2025-12-16 10:57:08 +0100
  • c2f3f7a20e ggml : use WARP_SIZE/2 for argmax reduction offset Aadeshveer Singh 2025-12-16 17:00:18 +0530
  • 4a78ebac82 fix and tested LLM_ARCH_NEMOTRON_H_MOE Xuan-Son Nguyen 2025-12-16 12:23:51 +0100
  • dffb032ca8 show more meaningful error message on missing tensor Xuan-Son Nguyen 2025-12-16 12:22:48 +0100
  • 51c3de6887 Merge remote-tracking branch 'sfallah/master' into sf/deepseek-ocr Saba Fallah 2025-12-16 12:16:25 +0100
  • f4b088c5fa fix LLM_ARCH_NEMOTRON_H_MOE Xuan Son Nguyen 2025-12-16 12:06:00 +0100
  • 942ddbe900 Merge branch 'master' into xsn/arch_refactor_llm_names Xuan Son Nguyen 2025-12-16 12:03:21 +0100
  • 7b1db3d3b7
         arg: clarify auto kvu/np being set on server (#17997) b7433 Xuan-Son Nguyen 2025-12-16 12:01:27 +0100
  • a5251ca11d
         Optimization: Qwen3 next autoregressive pass (#17996) b7432 Piotr Wilkin (ilintar) 2025-12-16 11:59:53 +0100
  • fb644247de
         CLI: fixed adding cli and completion into docker containers, improved docs (#18003) Andrew Aladjev 2025-12-16 13:52:23 +0300
  • 5f5f9b4637
         server: Update README.md incorrect argument (#18073) 2114L3 2025-12-16 20:50:43 +1000
  • ea773421e8 chore: update webui static output Aleksander Grygier 2025-12-16 11:36:11 +0100
  • c16d12de27 refactor: Clean up & add JSDocs Aleksander Grygier 2025-12-16 11:18:44 +0100
  • 3d86c6c2b5
         model: support GLM4V vision encoder (#18042) b7429 Xuan-Son Nguyen 2025-12-16 11:25:26 +0100
  • 9963b81f63
         model-conversion : add note about verifying previous models (#18082) Daniel Bevenius 2025-12-16 11:17:40 +0100
  • db81d5ec4b
         model-conversion : use CONVERTED_EMBEDDING_MODEL for embedding_verify_logits (#18079) Daniel Bevenius 2025-12-16 11:17:20 +0100
  • ef852f6d5f
         Merge 042b4947f1 into c05aa69f32 Raul Torres 2025-12-16 10:13:34 +0000
  • 9706f77e6e
         Merge 976dbbd363 into c05aa69f32 Raul Torres 2025-12-16 10:13:18 +0000
  • c05aa69f32
         common : add nemotron 3 parsing (#18077) b7426 Aldehir Rojas 2025-12-16 04:05:23 -0600
  • 279cef27c2
         added note for old Intel hardware pre sycl (#18017) Francisco Herrera 2025-12-16 04:45:09 -0500
  • 179a09e3ac fix: Remove runes Aleksander Grygier 2025-12-16 10:32:21 +0100
  • 52c283a951 server : add encoder-decoder model support (T5, BART, MADLAD) Turkka Mannila 2025-12-12 11:47:23 +0200
  • 3a9bce2b1e refactor: ID generation improvements Aleksander Grygier 2025-12-16 10:24:00 +0100
  • 5ba95754ee
         security : add collaborator guidance (#18081) Georgi Gerganov 2025-12-16 11:17:11 +0200
  • 0e7e5db661 refactor: DRY Markdown post-processing logic Aleksander Grygier 2025-12-16 10:14:36 +0100
  • 8fa720606a
         ggml: migrate work_data to stack allocation Herman Semenoff 2025-12-16 12:01:39 +0300
  • 121e192865
         server: add optional POST /exit endpoint for graceful shutdown Akarshan 2025-12-16 14:25:26 +0530
  • ad1b60abc4
         Merge remote-tracking branch 'upstream/master' into backend-sampling Daniel Bevenius 2025-12-16 09:45:08 +0100
  • 755aeef41c
         Merge d3aea508a1 into 2aa45ef9e3 h9-tec 2025-12-16 16:38:07 +0800
  • cb1f80704c
         model-conversion : add note about verifying previous models Daniel Bevenius 2025-12-16 09:30:52 +0100
  • c05da389ba refactor: Logic improvements Aleksander Grygier 2025-12-16 09:29:38 +0100
  • 3c8b4f5203
         Merge c25eb6f7c5 into 2aa45ef9e3 Yihao Wang 2025-12-16 16:23:07 +0800
  • e47a082fc9
         security : add collaborator guidance gg/security-update Georgi Gerganov 2025-12-16 10:16:46 +0200
  • 4a4f82968c
         Merge branch 'ggml-org:master' into sf/deepseek-ocr Saba Fallah 2025-12-16 09:09:52 +0100
  • 0019f6e347
         model-conversion : use CONVERTED_EMBEDDING_MODEL for embedding_verify_logits Daniel Bevenius 2025-12-16 09:06:31 +0100
  • 9d0bcfaf4b chore: add index.html.gz Kim Simonsen 2025-12-13 01:37:19 +0100
  • 5979b853b5 webui: fix chat header width when sidebar is closed Kim Simonsen 2025-12-13 01:19:01 +0100
  • 486d29ee3d
         Merge e516cd0056 into 2aa45ef9e3 Uttam Pawar 2025-12-16 13:24:29 +0530
  • 2aa45ef9e3
         llama: Include algorithm header needed for C++23 (#18078) b7423 Chris Peterson 2025-12-15 23:37:55 -0800
  • 5362340dca llama: fix early stop in params_fit if ctx is set Johannes Gäßler 2025-12-15 23:17:23 +0100
  • c560316440
         graph : reuse SSM graphs (#16490) b7422 Georgi Gerganov 2025-12-16 09:36:21 +0200
  • bb4dd82e9e draft: incremental markdown rendering with stable blocks Pascal 2025-12-09 21:50:01 +0100
  • d6742125c3
         ci : separate webui from server (#18072) Sigbjørn Skjæret 2025-12-16 08:17:26 +0100
  • 8848d5df15 llama: Include algorithm header needed for C++23 Chris Peterson 2025-12-15 22:52:58 -0800
  • d616fee7d9 remove debug line Alde Rojas 2025-12-16 00:49:32 -0600
  • 3034836d36
         webui: Improve copy to clipboard with text attachments (#17969) Aleksander Grygier 2025-12-16 07:38:46 +0100
  • 6dce495096 chore: update webui static output Aleksander Grygier 2025-12-16 07:35:13 +0100
  • 1eab64f32c chore: update webui build output Aleksander Grygier 2025-12-15 12:18:41 +0100
  • 5ad78fcced fix: Decode HTML entities using `DOMParser` Aleksander Grygier 2025-12-15 12:14:43 +0100
  • dc736cd5b0 chore: update webui static output Aleksander Grygier 2025-12-12 18:07:23 +0100
  • 0bdb3a8d34 fix: UI issues Aleksander Grygier 2025-12-12 17:21:40 +0100
  • cab1d426fc chore: update webui static output Aleksander Grygier 2025-12-12 15:37:53 +0100
  • b5d2587b15 chore: update webui build output Aleksander Grygier 2025-12-12 14:19:41 +0100
  • 6ac4526c96 feat: Create copy/paste user message including "pasted text" attachments Aleksander Grygier 2025-12-12 13:54:43 +0100