Public/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-19 09:15:04 +02:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 3d7baaebb4 Merge remote-tracking branch 'origin/master' into qwen4exp/qwen3.8-flash-next pr-27742-tmp Xuan Son Nguyen 2026-08-27 18:32:21 +02:00
  • 732707dff2 quantize: cap working memory size to avoid loading big tensors onto RAM (#27795) b10656 Xuan-Son Nguyen 2026-08-27 18:31:13 +02:00
  • a510c82e17 LazyChunkedTensor Xuan Son Nguyen 2026-08-27 18:29:28 +02:00
  • fbc5135abb improve the m-rope image for get_prev_tokens Xuan Son Nguyen 2026-08-27 17:35:12 +02:00
  • cb300598d5 Feature: Added LIGHTNING_INDEXER support for Deepseek V4 ops on Vulkan Backend (#27453) b10655 Shobhit 2026-08-27 21:34:42 +08:00
  • 250b61446e replace some changes with #27795 Xuan Son Nguyen 2026-08-27 15:34:03 +02:00
  • 1a946ec745 pr2wt : use ssh/https remote in worktree depending on base (#27800) Sigbjørn Skjæret 2026-08-27 15:27:17 +02:00
  • af1ffaf37f nits Xuan Son Nguyen 2026-08-27 15:21:59 +02:00
  • 539c35e940 Merge remote-tracking branch 'origin/master' into qwen4exp/qwen3.8-flash-next Xuan Son Nguyen 2026-08-27 15:19:24 +02:00
  • fac889fb38 llama: model_loader: add TENSOR_READ_LAZY (#27794) b10653 Xuan-Son Nguyen 2026-08-27 15:14:34 +02:00
  • 41b4b7073c revert LLAMA_MMAP_RANDOM Xuan Son Nguyen 2026-08-27 15:01:35 +02:00
  • cae63579b6 ui: Improve Chat Form Actions UI/UX (models selector, add panel) (#27746) Aleksander Grygier 2026-08-27 14:47:36 +02:00
  • 243914706f clean up new comments Xuan Son Nguyen 2026-08-27 14:43:04 +02:00
  • 56672bf6dd Merge commit 'refs/pull/27742/head' of github.com:ggerganov/llama.cpp into qwen4exp/qwen3.8-flash-next Xuan Son Nguyen 2026-08-27 14:41:39 +02:00
  • bcb6084a4e convert : fix Nemotron-H LoRA GGUF conversion (#27356) Kartik Gulia 2026-08-27 18:11:24 +05:30
  • 5b34ce297f clean up code comments Xuan Son Nguyen 2026-08-27 14:34:48 +02:00
  • 937fe54d11 address comments xsn/tensor_rows_lazy Xuan Son Nguyen 2026-08-27 13:55:55 +02:00
  • c0c7fa930d quantize: cap working memory size to avoid loading big tensors onto RAM xsn/quant_cap Xuan Son Nguyen 2026-08-27 13:48:24 +02:00
  • ef6876693f qwen4exp: bias the QSA selection per block, not per cell danielhanchen 2026-08-27 10:40:50 +00:00
  • 420c032e9f gen docs Xuan Son Nguyen 2026-08-27 13:29:30 +02:00
  • e5227ab28e FACP (Fewer Acronym Classes Please) Sigbjørn Skjæret 2026-08-27 12:15:18 +02:00
  • f3bb7618c5 rename to TENSOR_READ_LAZY Xuan Son Nguyen 2026-08-27 13:24:29 +02:00
  • bb07e28f88 add --tensor-read-lazy Xuan Son Nguyen 2026-08-27 13:10:03 +02:00
  • db97713d10 ui: split model option icons into capabilities and modalities allozaur/feat/chat-form-actions/model-selector-ux Aleksander Grygier 2026-08-26 17:27:51 +02:00
  • 6860acec69 ui: derive modality badge icons and labels from shared constants Aleksander Grygier 2026-08-26 17:27:37 +02:00
  • 30e99fb44b ui: add ModelCapability enum and shared modality/capability icon constants Aleksander Grygier 2026-08-26 17:27:28 +02:00
  • 70f98ee650 chore: Format Aleksander Grygier 2026-08-26 15:35:50 +02:00
  • edd20addad chore: Format Aleksander Grygier 2026-08-26 15:33:43 +02:00
  • d3f636b8e8 ui: show modality icons instead of file submenu in chat add menu Aleksander Grygier 2026-08-22 13:11:56 +02:00
  • 9ae40a18c3 ui : fix eslint issues in chat form and model selector Aleksander Grygier 2026-08-17 11:04:02 +02:00
  • 1e64786209 feat: Enable microphone input as default for audio models Aleksander Grygier 2026-08-16 18:03:52 +02:00
  • 0daae104c5 ui : add raw model id tooltip to model selector options Aleksander Grygier 2026-08-15 23:26:25 +02:00
  • c651343a00 ui : make model option hover and focus highlight override the active state Aleksander Grygier 2026-08-15 22:54:05 +02:00
  • 9e0e0e31ed ui : move model list into a submenu within the model selector Aleksander Grygier 2026-08-15 22:53:56 +02:00
  • 4d87d4e275 ui : add show-org-name-in-trigger display setting Aleksander Grygier 2026-08-15 22:53:48 +02:00
  • f2612f5389 ui : keep reasoning submenu visible regardless of model state Aleksander Grygier 2026-08-15 22:53:26 +02:00
  • a8826afdf3 ui : show reasoning and modality icons on model options and search by modality Aleksander Grygier 2026-08-15 22:53:18 +02:00
  • 01a0f390c3 ui : strip trailing container-format segments from parsed model names Aleksander Grygier 2026-08-15 22:53:04 +02:00
  • fe235f4343 ui: Replace per-conversation MCP overrides with per-conversation tool policy (#27745) Aleksander Grygier 2026-08-27 13:08:01 +02:00
  • e072597f9f llama: model_loader: add TENSOR_GET_ROW_LAZY Xuan Son Nguyen 2026-08-27 12:54:09 +02:00
  • 2bb9bddafa spec: Add benchmark-only synthetic speculative acceptance options (#27711) b10649 Gaurav Garg 2026-08-27 16:23:42 +05:30
  • deae5ee133 model : simplify MiniMax-01 graph (#27790) b10648 fairydreaming 2026-08-27 12:27:52 +02:00
  • 69a1269f70 Merge commit 'refs/pull/27742/head' of github.com:ggerganov/llama.cpp into qwen4exp/qwen3.8-flash-next Xuan Son Nguyen 2026-08-27 12:13:47 +02:00
  • f29551215b args: add --video-* CLI arguments (#24318) b10647 Xuan-Son Nguyen 2026-08-27 12:11:12 +02:00
  • b8bdf73bb9 llama: fold the random-access prefetch into its own feature flag Daniel Han 2026-08-27 09:07:02 +00:00
  • 58325573c9 llama: narrow the random-access mmap advice to the gather table danielhanchen 2026-08-27 06:06:07 +00:00
  • 95da4ba860 llama: opt-in random-access mmap advice for host-resident gather tables danielhanchen 2026-08-27 04:20:48 +00:00
  • c3259772c7 qwen4exp: shrink the PLE hparams storage danielhanchen 2026-08-27 07:32:16 +00:00
  • 1caf6c5711 llama: give the qwen4exp full memory context its indexer cache Daniel Han 2026-08-27 06:16:54 +00:00
  • 5eeb97d1e4 no more ple_hist (use master version) Xuan Son Nguyen 2026-08-27 11:57:52 +02:00
  • b5b5fc30b3 Merge remote-tracking branch 'origin/master' into qwen4exp/qwen3.8-flash-next Xuan Son Nguyen 2026-08-27 11:57:27 +02:00
  • 915dc6d38c metal : fix memory leaks due to missing autoreleasepools (#27758) b10646 Niklas Wenzel 2026-08-27 11:53:08 +02:00
  • 4e69771682 add mtmd_helper_init_opt xsn/video_args Xuan Son Nguyen 2026-08-27 11:27:48 +02:00
  • c5fc7e3488 llama : add --n-cpu-ffn option (#26622) b10645 Jonas J 2026-08-27 12:26:42 +03:00
  • 0cfa2beec7 ui: remove dead MCP resources menu trigger chain allozaur/feat/chat-form-actions/tool-policy Aleksander Grygier 2026-08-27 11:20:55 +02:00
  • 3d964e84f2 ui: keep MCP connections stable across policy switches Aleksander Grygier 2026-08-27 11:19:21 +02:00
  • 9b521a7caf fix: UI for Model Information dialog Aleksander Grygier 2026-08-27 11:00:27 +02:00
  • 85acaa0626 ui: fix group checkbox sticking checked after disable Aleksander Grygier 2026-08-27 10:38:13 +02:00
  • cf39a070ad ui: render dash for mixed-state group checkboxes Aleksander Grygier 2026-08-27 10:18:01 +02:00
  • dc2dd9913c ui: remove dead MCP prompt menu trigger chain Aleksander Grygier 2026-08-27 10:15:42 +02:00
  • 2e8cce869e ui: gate /prompt command on the conversation tool policy Aleksander Grygier 2026-08-27 10:04:55 +02:00
  • e415366eba ui: derive group checkbox state in useToolsPanel Aleksander Grygier 2026-08-27 09:54:09 +02:00
  • cbace8af70 ui: initialize only policy-enabled MCP servers for a flow Aleksander Grygier 2026-08-27 09:51:08 +02:00
  • b520daf9a7 ui: let indeterminate state win over checked in group checkboxes Aleksander Grygier 2026-08-27 09:50:58 +02:00
  • 20b0733cee ui: replace literal comment text in sheet group snippet Aleksander Grygier 2026-08-27 09:50:52 +02:00
  • 9bb932d16d ui: scroll model information header with the content Aleksander Grygier 2026-08-27 08:39:35 +02:00
  • e88da90c30 ui: stack chat template row in model information dialog Aleksander Grygier 2026-08-27 08:23:03 +02:00
  • b76386d6e2 ui: make model information dialog full-screen on mobile Aleksander Grygier 2026-08-27 08:22:45 +02:00
  • 116599b1b7 ui: use fixed table layout in model information dialog Aleksander Grygier 2026-08-27 08:17:57 +02:00
  • 075dd71592 ui: scroll wide chat template in model information dialog Aleksander Grygier 2026-08-27 08:11:17 +02:00
  • a72d3f12c8 ui: fix model information dialog width on all screen sizes Aleksander Grygier 2026-08-27 08:06:25 +02:00
  • e87eff0350 ui: remove unmounted MCP submenu component Aleksander Grygier 2026-08-27 07:29:32 +02:00
  • cd9d02c63f ui: gate MCP prompt and resource capabilities on tool policy Aleksander Grygier 2026-08-27 07:29:21 +02:00
  • ba419bf2f1 ui: indeterminate group checkboxes and inert grayed rows Aleksander Grygier 2026-08-27 06:44:44 +02:00
  • 4f2f679559 ui: clean up tool key helpers and store docs Aleksander Grygier 2026-08-27 06:44:24 +02:00
  • f6346a6ae1 ui: gate cwd picker and mention picker on effective tool policy Aleksander Grygier 2026-08-27 06:44:12 +02:00
  • a9ea5b9f3d ui: make Settings tools tab edit defaults with category toggles Aleksander Grygier 2026-08-27 06:44:01 +02:00
  • 0de27be4fa ui: skip MCP init when no policy-enabled server remains Aleksander Grygier 2026-08-27 06:43:53 +02:00
  • 473d4a3faf ui: clear MCP server group key in enableAllToolsForServer Aleksander Grygier 2026-08-27 06:43:46 +02:00
  • f51b3cd95c ui: restore reasoning section in mobile add sheet Aleksander Grygier 2026-08-27 06:35:01 +02:00
  • f1a9607d0c chore: format Aleksander Grygier 2026-08-26 21:00:18 +02:00
  • 617d9aed96 ui: skip MCP init when flow policy disables the MCP category Aleksander Grygier 2026-08-26 20:48:34 +02:00
  • 9eb3f7df95 ui: drop unused MCP category toggle from tools panel hook Aleksander Grygier 2026-08-26 20:48:28 +02:00
  • d018525528 ui: gate MCP server avatars on conversation tool policy Aleksander Grygier 2026-08-26 20:34:28 +02:00
  • c7dd200b37 ui: align preferences section headers with their methods Aleksander Grygier 2026-08-26 20:34:23 +02:00
  • 40d590623e ui: fall back to global defaults when agentic flow has no tool policy Aleksander Grygier 2026-08-26 20:34:06 +02:00
  • 38c4721fbd ui: keep tool policy migration running when defaults parse fails Aleksander Grygier 2026-08-26 20:34:00 +02:00
  • c8d5a69767 ui: replace per-conversation MCP overrides with per-conversation tool policy Aleksander Grygier 2026-08-22 08:03:42 +02:00
  • bf5396021e qwen4exp: give the PLE conv history its own mirrored recurrent row Daniel Han 2026-08-27 04:42:36 +00:00
  • d1fe507f95 kv-cache: check the mirrored slot layout on a whole-context restore too Daniel Han 2026-08-27 04:13:22 +00:00
  • f43070aa5c kv-cache: clear the cache once when restoring a whole context Daniel Han 2026-08-27 04:13:14 +00:00
  • 8b3ed0a400 qwen4exp: keep the indexer cache in step across server slots Daniel Han 2026-08-27 02:35:16 +00:00
  • d7a2074112 models : support nanbeige4.2-3B (#27730) b10644 zql 2026-08-27 12:55:31 +08:00
  • 6c5afc86ae qwen4exp: double the Q split granularity for tensor parallelism danielhanchen 2026-08-27 03:47:07 +00:00
  • fbe17732bd llama: give the qwen4exp indexer cache its own tensor names danielhanchen 2026-08-27 03:41:50 +00:00
  • 213df585b9 convert: fix the qwen4exp lint and type-check failures Daniel Han 2026-08-27 02:26:31 +00:00
  • 192067b72d hexagon: support for multi-NPU devices (IQ9, IQ10) and fully asynchronous backend (#26501) b10643 Max Krasnyansky 2026-08-26 18:46:50 -07:00
  • 0b19188e93 qwen4exp: include llama-impl.h explicitly for llama_mul_mat_hadamard danielhanchen 2026-08-27 01:09:27 +00:00
  • 24ea62df44 llama: fix the qwen4exp PLE history seq_rm(-1) iterator invalidation and the fatal-warning build Daniel Han 2026-08-27 01:18:11 +00:00
  • 5674c73aa8 llama: segment the qwen4exp fused QKV for tensor split Daniel Han 2026-08-27 01:14:17 +00:00