mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 31 Packages Projects Releases Wiki Activity

Commit Graph

  • b34ac655f3 merge: TurboQuant KV cache types onto upstream master (MTP built-in) mtp-turboquant root 2026-05-18 01:52:24 +03:00
  • 726704a160 feat: Support d_conv=15 for ssm-conv.cu (#23017) Gabe Goodhart 2026-05-17 15:05:11 -06:00
  • 87589042ca cmake : fix LLAMA_BUILD_UI logic (#23190) Aldehir Rojas 2026-05-17 14:42:26 -04:00
  • e0de4c2419 cmake : do not install conversion script (#23204) Sigbjørn Skjæret 2026-05-17 18:07:21 +02:00
  • 84c678242a CUDA: Continue directly including cuda/iterator (#23102) Oliver Simons 2026-05-17 18:00:10 +02:00
  • 3e12fbdea5 llama: avoid copying logits during prompt decode in MTP (#23198) Aman Gupta 2026-05-17 23:30:25 +08:00
  • 39cf5d6191 common : delegate assistant continuation to underlying template handlers (#23089) Aldehir Rojas 2026-05-17 07:36:05 -04:00
  • a6d6183dbc ggml-vulkan/CMakeLists: add a check for SPIRV-Headers (#22009) Jan Ekström 2026-05-17 14:12:11 +03:00
  • fcae601e44 vulkan: add cpy bf16 -> f32 pipelines (#22677) Pascal 2026-05-17 11:31:20 +02:00
  • 7ba22c6a09 vulkan: Support unaligned tensors for ROPE (#22637) Jeff Bolz 2026-05-17 04:30:16 -05:00
  • f4cc787b9f common : enable streaming JSON argument values (#23173) Aldehir Rojas 2026-05-17 04:44:34 -04:00
  • 3fbadb06dc vulkan: fuse SSM_CONV + BIAS + SILU (#22653) Jeff Bolz 2026-05-17 03:25:50 -05:00
  • 1a68ec9378 server : honor --embd-normalize CLI arg (#23125) Rares Vernica 2026-05-16 23:39:04 -07:00
  • a16cce81d3 ngram : reduce noisy logs (#23185) ddh0 2026-05-17 01:38:17 -05:00
  • 4f13cb7424 webui: support video files as input (#22830) Judd 2026-05-17 08:13:44 +08:00
  • b64739ea39 server: (router) alloc tmp buffer on heap (#23159) Xuan-Son Nguyen 2026-05-16 23:42:16 +02:00
  • 64b38b561b server: skip device enumeration in router mode to avoid creating CUDA primary context (#23137) Pascal 2026-05-16 21:21:06 +02:00
  • 6049906133 vulkan: removed duplicate #include <memory> in headers (#23144) Winston Ma 2026-05-17 01:57:35 +08:00
  • 0253fb21f5 ui: Add request timeout for MCP tool calls (#23138) Aleksander Grygier 2026-05-16 15:20:27 +02:00
  • 3a92bc99db sync : ggml Georgi Gerganov 2026-05-16 15:59:45 +03:00
  • e6c37a1adc ggml : bump version to 0.12.0 (ggml/1494) Georgi Gerganov 2026-05-16 15:59:09 +03:00
  • 560445bf34 metal : tighten input-position loop in kernel_conv_transpose_1d (ggml/1477) CrispStrobe 2026-05-10 16:45:00 +02:00
  • 2eb3e6b242 ggml: install ggml.pc in <libdir>/pkgconfig (ggml/1480) Steve Lhomme 2026-05-10 16:35:38 +02:00
  • 25b1bc9c2f ui: Correct links in tools/ui/README.md [no ci] (#23139) Holger Voormann 2026-05-16 14:42:38 +02:00
  • 18675b6bbc vendor : update cpp-httplib to 0.45.0 (#23103) Alessandro de Oliveira Faria (A.K.A.CABELO) 2026-05-16 09:25:21 -03:00
  • 255582687b llama + spec: MTP Support (#22673) Aman Gupta 2026-05-16 20:06:23 +08:00
  • b81c2cdd74 ui: Fix handling of MCP resource template parameters (#23117) kubawoo 2026-05-16 13:25:41 +02:00
  • 1428004808 webui : [ChatFormActionAdd][a11y] fix accessibility issues in add menu trigger and items (#22736) viggy 2026-05-16 03:00:46 -07:00
  • 366c5e2a3b ui: untrack settings sync in props effect to prevent reactive loop (#23127) Pascal 2026-05-16 11:25:34 +02:00
  • 1d9f99aa75 fix: Add build step using build workflow to publish workflow (#23134) Aleksander Grygier 2026-05-16 11:22:59 +02:00
  • 42928bc14d model : NvFP4 quantized LM head support (#23046) ynankani 2026-05-16 09:09:27 +00:00
  • 59778f0196 ui: Restructure repo to use tools/ui folder and ui / UI / llama-ui / LLAMA_UI naming (#23064) Aleksander Grygier 2026-05-16 02:02:40 +02:00
  • 49d1701bd2 ci : fix release symlinks (#23119) Sigbjørn Skjæret 2026-05-16 01:09:28 +02:00
  • 1348f67c58 webui: Use lowercase hash for HF checksum check (#23107) Omer Ozarslan 2026-05-15 10:38:16 -07:00
  • cfabeb1bad tests: add BF16 non-contig coverage for MUL_MAT permutations (#22689) Pascal 2026-05-15 19:35:05 +02:00
  • 6831fe470c docs: document usage object in server timings response (#23110) Julien Chaumond 2026-05-15 19:33:12 +02:00
  • 72e60f500d mtmd: add chunks and fix preproc for qwen3a (#23073) Xuan-Son Nguyen 2026-05-15 19:32:47 +02:00
  • 8be1786707 webui: fix theme from --webui-config-file not applied on first load (fresh localStorage) (#22902) Pascal 2026-05-15 19:25:38 +02:00
  • 18d1717d62 convert : fix Qwen3 ASR conversion (#23081) Sigbjørn Skjæret 2026-05-15 18:38:39 +02:00
  • cc7200bf12 Refactor: convert_hf_to_gguf.py (#17114) Piotr Wilkin (ilintar) 2026-05-15 15:18:12 +02:00
  • 769cc93a43 ci : fix transform of top . entry in release archive (#23080) Sigbjørn Skjæret 2026-05-15 13:13:16 +02:00
  • d5dc2e0a02 llama-eval : add AIME 2026 dataset support (#23058) Georgi Gerganov 2026-05-15 13:58:30 +03:00
  • ac33f032ac reasoning-budget: clone should do a deep-copy (#23095) Aman Gupta 2026-05-15 17:59:07 +08:00
  • d528444580 webui: preserve partial response on streaming error (#23090) Pascal 2026-05-15 11:18:11 +02:00
  • 91e84fed64 Support for Codex CLI by skipping unsupported Responses tools (#23041) Sid Shaytay 2026-05-15 00:03:24 -07:00
  • 7155a49771 readme : update bindings (#23063) KITAITI Makoto 2026-05-15 14:41:24 +09:00
  • 5c0e946837 ggml-hexagon: cpy: add contiguous fast-path in reshape copy (#23076) Pranav Dhinakar 2026-05-14 16:55:54 -07:00
  • 3e037f313c HIP: RDNA3 mma FA, faster AMD transpose, tune AMD (#22880) Johannes Gäßler 2026-05-14 22:58:58 +02:00
  • d81e63dcfd CI : support IOT device (IQ9) (#22987) Zack Li 2026-05-14 13:58:34 -07:00
  • 834a243664 ggml-webgpu: Enable NVIDIA self-hosted CI (#22976) Reese Levine 2026-05-14 09:41:32 -07:00
  • 5ec717d125 ggml-webgpu: makes the flash attn vec path subgroup-aware (#23040) Zheyuan Chen 2026-05-14 09:31:36 -07:00
  • 0c3e4fccca fix: Propagate version tag to WebUI asset download in self-hosted CI (#23051) Aleksander Grygier 2026-05-14 17:57:20 +02:00
  • 97b658cee8 contributing: new contributors should not submit trivial fixes (#23045) Aman Gupta 2026-05-14 23:55:24 +08:00
  • 253ba110bc webui: Move static build output from repo code to HF Bucket (#22937) Aleksander Grygier 2026-05-14 13:21:41 +02:00
  • 67b2b7f2f2 logs : reduce (#23021) Georgi Gerganov 2026-05-14 13:05:52 +03:00
  • 81b0d882ae ggml-cpu: Add IME2 Instruction Support for the SpacemiT Backend (#22863) alex-spacemit 2026-05-14 17:39:30 +08:00
  • 0f45f1a35c docker : revert stable version of intel compute-runtime (#22968) Neo Zhang 2026-05-14 17:30:40 +08:00
  • 42532afff4 unicode,test: add Qwen3.5 non-backtracking tokenizer handler and regr… (#22110) Kabir Potdar 2026-05-14 09:03:40 +00:00
  • dbe7901ca6 vulkan: fix matmul integer pipeline selection (#23005) Ruben Ortlam 2026-05-14 10:36:54 +02:00
  • 320a6a44a5 fix: Autoscroll detection (#23026) Aleksander Grygier 2026-05-14 08:09:29 +02:00
  • 9ed6e19b9d SYCL: fix multi-GPU system RAM exhaustion by using Level Zero allocations (#21597) Katostrofik 2026-05-14 01:39:14 -04:00
  • 4c1c3ac09d ggml-webgpu: only use subgroup-matrix path when head dims are divisible by sg_mat_k / sg_mat_n (#23020) Zheyuan Chen 2026-05-13 15:12:40 -07:00
  • 7f3f843c31 Fix for issue #22974. Cast intermediate results to float before adding and casting the result to the destination type. Avoids half+half operator ambiguity. (#22994) scutler-nv 2026-05-13 13:36:14 -07:00
  • ec562eb673 opencl: add q5_0 and q5_1 MoE for Adreno (#22985) shaofeiqi 2026-05-13 11:57:31 -07:00
  • 95d469a915 server, webui: accept continue_final_message flag for vLLM API compat (#23012) Pascal 2026-05-13 20:47:58 +02:00
  • 1e4579fbb8 opencl: fix crash when warming up MoE on Adreno (#22876) lhez 2026-05-13 11:24:33 -07:00
  • 527045bfb0 flush the gpu profile timestamp before the queryset is overflowed (#22995) Masashi Yoshimura 2026-05-14 02:22:44 +09:00
  • 2dfeca31cc webui: Deduplicate model aliases in data + handle single/multiple aliases in UI (#22979) Aleksander Grygier 2026-05-13 16:39:36 +02:00
  • 46be24d121 webui: preserve system message on edit cancel (#22911) Pascal 2026-05-13 16:16:02 +02:00
  • 7e16646015 docs : Update OPENVINO.md (#22959) Ravi Panchumarthy 2026-05-13 07:12:15 -07:00
  • ad96bb8c0c hexagon: add unary tanh op (#22999) Max Krasnyansky 2026-05-13 06:59:28 -07:00
  • e75cd5efb5 download: do not exit() on error (#23008) Xuan-Son Nguyen 2026-05-13 15:14:58 +02:00
  • 5d44db6008 server, webui: support continue generation on reasoning models (#22727) Pascal 2026-05-13 11:09:51 +02:00
  • 3796c94bad ci: validate model naming convention (#22680) Xuan-Son Nguyen 2026-05-13 10:59:37 +02:00
  • 634275fbbb spec : update CLI arguments for better consistency (#22964) Georgi Gerganov 2026-05-13 09:15:39 +03:00
  • bcfe63fc53 llama-eval : enable type check (#22988) Sigbjørn Skjæret 2026-05-13 08:14:24 +02:00
  • 61af07c22d ggml-zendnn : adaptive fallback to CPU backend for small batch sizes (#22681) Sachin Sharma 2026-05-13 11:43:47 +05:30
  • 856c3adac1 hexagon: eliminate scalar VTCM loads via HVX splat helpers (#22993) Trivikram Reddy 2026-05-12 19:28:02 -05:00
  • a9883db8ee opencl: add opt-in Adreno xmem F16xF32 GEMM for prefill (#22755) yzyyzyhhh 2026-05-13 04:10:37 +08:00
  • cce09f0b2b convert : fix Pixtral 12B --mistral-format conversion (3 bugs) (#22981) fredzillman 2026-05-12 17:16:01 -02:30
  • dded58b450 webui: Fix Chat Screen Form box disappearing + autoscroll issues on WebKit (#22977) Aleksander Grygier 2026-05-12 20:41:11 +02:00
  • 7bfe120c21 mtmd, server, common: expose modalities to /v1/models (#22952) Xuan-Son Nguyen 2026-05-12 19:08:07 +02:00
  • 927dada6c9 ggml-webgpu: Enables running gpt-oss-20b (#22906) Masashi Yoshimura 2026-05-12 23:27:40 +09:00
  • 239a497e5f ggml-webgpu: address precision issues for multimodal (#22808) Chen Yuan 2026-05-12 10:27:04 -04:00
  • 89730c8d26 model-conversion : add causal-convert-mmproj target [no ci] (#22969) Daniel Bevenius 2026-05-12 15:15:40 +02:00
  • fde69a3607 examples : add llama-eval (#21152) Georgi Gerganov 2026-05-12 15:07:00 +03:00
  • ef93e98d01 vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461) Masato Nakasaka 2026-05-12 03:15:34 -07:00
  • 706fbd8ab6 vulkan: Check shared memory size for mmq shaders (#22693) Jeff Bolz 2026-05-12 04:41:58 -05:00
  • fa62042af9 ci : bump ty to 0.0.35 (#22961) Sigbjørn Skjæret 2026-05-12 11:34:10 +02:00
  • 4178259130 mtmd: add MiMo v2.5 vision (#22883) AesSedai 2026-05-12 02:11:14 -07:00
  • 78fbbc2c07 convert : add split() to LoraTorchTensor in LoRA converter (#22832) Jesus Talavera 2026-05-12 07:17:04 +02:00
  • da44953329 metal : promote mul_mv/mul_mm batch divisors to function constants (#22711) guyfischman 2026-05-12 07:15:02 +02:00
  • 1ec7ba0c14 opencl: add q4_1 MoE for Adreno (#22856) Shawn Gu 2026-05-11 11:57:26 -07:00
  • 8e1f9d0834 CUDA: handle OW > 65535 in im2col (2D and 3D) (#22944) CrispStrobe 2026-05-11 19:48:29 +02:00
  • e936660760 Ggml/cuda snake fusion hardening (#22912) Pascal 2026-05-11 18:42:08 +02:00
  • ef22b3e4ac docs: fix metrics endpoint description in server README (#22879) willjoha 2026-05-11 18:32:26 +02:00
  • 68e7ea3eab spec : parallel drafting support (#22838) Georgi Gerganov 2026-05-11 19:09:43 +03:00
  • 928b486b0c ggml-virtgpu: Add a GHA build check (#22943) Kevin Pouget 2026-05-11 15:38:22 +02:00
  • 7dbb0e998a examples : update args speculative-simple README.md [no ci] (#22938) Daniel Bevenius 2026-05-11 13:00:57 +02:00
  • dd9280a664 vulkan: Support asymmetric FA in scalar/mmq/coopmat1 paths (#22589) Jeff Bolz 2026-05-11 05:49:03 -05:00