]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2026-05-04 Charles Xukleidiai : update to v1.24.0 and use release archive...
2026-05-04 leonardHONGCUDA: use fastdiv for batch index split in get_rows...
2026-05-04 Xuan-Son Nguyenserver: implement /models?reload=1 (#21848)
2026-05-04 Shakhnazar... examples: refactor diffusion generation (#22590)
2026-05-04 JusteLeowebui : fix circular dependency between chat.service...
2026-05-04 Piotr Wilkin... common/autoparser: fixes for newline handling / forced...
2026-05-04 Xuan-Son Nguyenmodel: move `load_hparams` and `load_tensors` to per...
2026-05-04 Evan Huusserver: Add a simple get_datetime server tool (#22649)
2026-05-04 Nick Towlewebui: restore missing settings (#22666)
2026-05-04 Georgi Gerganovdocs : update speculative decoding parameters after...
2026-05-04 Atomic-Germvulkan: delete dead GGML_VK_MAX_NODES def (#22621)
2026-05-04 Chen Yuanggml-webgpu: add layer norm ops (#22406)
2026-05-03 Aldehir Rojascommon : determine generation prompt using longest...
2026-05-03 Julien Denizeconvert : Mistral format yarn apply_scale support...
2026-05-03 JM Roblesconvert : apply Q/K RoPE permutation in NVFP4 repack...
2026-05-02 lucyfix: CUDA device PCI bus ID de-dupe OOMing (ignoring...
2026-05-02 Georgi Gerganovserver : avoid checkpoint data host copies (#22558)
2026-05-02 JusteLeoggml-virtgpu: fix circular dependency in headers (...
2026-05-02 Csaba Kecskemeticonvert : disable uint types (#18908)
2026-05-02 Shawn Guopencl: Adreno optimization for MoE - MxFP4 (#22301)
2026-05-02 Johannes GäßlerGithub: update issue templates (#22594)
2026-05-02 Georgi Gerganovsync : ggml
2026-05-02 Georgi Gerganovggml : bump version to 0.10.2 (ggml/1474)
2026-05-02 Georgi Gerganovsync : ggml
2026-05-02 Georgi Gerganovggml : try fix win32 build (whisper/0)
2026-05-02 Yiwei Shaohexagon: hmx flash attention (#22347)
2026-05-01 ddh0llama-quant : fix `--tensor-type` when default `qtype...
2026-05-01 Aparna M Phexagon: enable non-contiguous row tensor support for...
2026-05-01 Aleksander... webui: Spring Cleaning Refactor v1 (#22505)
2026-05-01 Masashi Yoshimuraggml-webgpu: Fix vectorized handling in mul-mat and...
2026-05-01 Jeff Bolzvulkan: Support asymmetric FA in coopmat2 path (#21753)
2026-05-01 Chen Yuanggml-webgpu: add the upscale shader (#22419)
2026-04-30 Masashi Yoshimuraggml-webgpu: Improve performance of mat-vec and mat...
2026-04-30 Reese LevineUpdate llama-mmap to use ftello/fseeko (#22497)
2026-04-30 Adrien Gallouëtcommon : check for null getpwuid in hf-cache (#22550)
2026-04-30 Ruben Ortlamvulkan: add get/set tensor 2d functions (#22514)
2026-04-30 Ben Guidarellispec: fix argument typo (#22552)
2026-04-30 Sigbjørn Skjæretci : bump ty to 0.0.33 (#22535)
2026-04-30 Adrien Gallouëtvendor : update cpp-httplib to 0.43.2 (#22548)
2026-04-30 Johannes GäßlerCUDA: fix tile FA kernel on Pascal (#22541)
2026-04-30 Georgi Gerganovscripts : add wc2wt.sh - create worktree from current...
2026-04-30 Rithik Sharmaadd fast matmul iquants (#22504)
2026-04-30 Georgi Gerganovspec : fix draft model checkpoints (#22521)
2026-04-30 Peter Siderisspec : fix vocab compat checks in spec example (#22426)
2026-04-29 Aldehir Rojascommon : do not pass prompt tokens to reasoning budget...
2026-04-29 Max Krasnyanskyhexagon: make vmem and buffer-size configurable (#22487)
2026-04-29 Anav PrasadCUDA: fuse SSM_CONV + ADD(bias) + SILU (#22478)
2026-04-29 Georgi Gerganovspec : disacard last drafted token with low prob (...
2026-04-29 Georgi Gerganovsync : ggml
2026-04-29 Georgi Gerganovggml : bump version to 0.10.1 (ggml/1469)
2026-04-29 Pascalwebui: fix slow mic stop and WAV encode (#22480)
2026-04-29 shalinib-ibmggml-cpu : disable tiled matmul on AIX to fix page...
2026-04-29 Aman Guptaggml-cuda: refactor fusion code (#22468)
2026-04-29 qiurui144ggml-cpu: cmake: append xsmtvdotii march for SpacemiT...
2026-04-29 Reese Levineggml-webgpu: Fix bug in FlashAttention support check...
2026-04-29 Masato Nakasakacommon: Intentionally leak logger instance to fix hangi...
2026-04-29 hrushitfujitsuggml : add sve tuned code for gemm_q8_0_4x8_q8_0()...
2026-04-29 Johannes GäßlerTP: fix delayed AllReduce + zero-sized slices (#22489)
2026-04-28 Michael Wandggml-cuda: Repost of 21896: Blackwell native NVFP4...
2026-04-28 lnigamggml-cuda: add flash-attn support for DKQ=320/DV=256...
2026-04-28 Daniel Beveniusconvert : add support for Nemotron Nano 3 Omni (#22481)
2026-04-28 Jillis ter... common : re-arm reasoning budget after DONE on new...
2026-04-28 Matt Corallovulkan: Coalesce Q4_K/Q5_K scale loads (#21751)
2026-04-28 Reese Levineggml-webgpu: fix buffer aliasing for ssm_scan and refac...
2026-04-28 Aleksander... webui: Server tools (#21237)
2026-04-28 Jeff Bolzvulkan: add barrier after writetimestamp (#21865)
2026-04-28 Emil Askerovggml: improve SPIR-V headers detection with __has_inclu...
2026-04-28 Adrien Gallouëtggml : skip already registered backends and devices...
2026-04-28 Adrien Gallouëtggml : revert to -lm linking instead of find_library...
2026-04-28 hipuddingCANN: add new ops, optimize existing ops (#21204)
2026-04-28 Georgi Gerganovspec : refactor params (#22397)
2026-04-28 Aman Guptaserver: use pos_next instead of n_tokens for m-rope...
2026-04-27 Rithik Sharmaggml-webgpu: add Q1_0 support (#22374)
2026-04-27 tha80server: (router) Forward form-data to model server...
2026-04-27 Rithik Sharmaadd fast mat-vec kernels for i-quants (#22344)
2026-04-27 Igor RudenkoAdditional test for common/gemma4 : handle parsing...
2026-04-27 unraidofix: rpc-server cache may not work in Windows environme...
2026-04-27 rankaiyxFix type casting for unaccounted memory calculation...
2026-04-27 Georgi Gerganovdownload : prefer q8_0 when q4_k not available (#22428)
2026-04-27 ynankanimodel : remove duplicate wo_s scale after build_attn...
2026-04-27 Sigbjørn Skjæretconvert : remove input_scale for dequantized fp8 modelo...
2026-04-27 Adrien Gallouëtggml : use 64 bytes aligned tile buffers (#21058)
2026-04-27 Max Krasnyanskycommon: fix missing exports in llama-common (#22340)
2026-04-26 Georgi Gerganovpr2wt : symlink .pi (#22386)
2026-04-26 Rithik Sharmaadd performance-portable tuning for register-tile and...
2026-04-26 Gaurav GargFix recurrent state serialization for partial reads...
2026-04-26 Johannes GäßlerGithub: set meta backend code owner (#22388)
2026-04-26 Oliver SimonsCUDA: better coalesce data-access for contiguous concat...
2026-04-26 Sigbjørn Skjæretggml-cpu : re-enable fast gelu_quick_f16 (#22339)
2026-04-26 Eveggml-cpu: optimize avx2 q6_k (#22345)
2026-04-26 lhezopencl: add iq4_nl support (#22272)
2026-04-26 Trivikram Reddyhexagon: guard HMX clock request for v75+ platforms...
2026-04-25 Piotr Wilkin... chat: fix handling of space in reasoning markers (...
2026-04-25 Georgi Gerganovspec : fix vocab compat checks (#22358)
2026-04-25 Johannes GäßlerCUDA: reduce MMQ stream-k overhead (#22298)
2026-04-25 Developer-Ecosystem... metal : optimize Metal Tensor API usage for GGML_OP_MUL...
2026-04-25 ddh0llama-quant : default ftype param `Q5_1` --> `Q8_0...
2026-04-25 Georgi Gerganovgitignore : add .pi + personal SYSTEM.md (#22316)
2026-04-25 Neo Zhang[SYCL] Optimize Q4_0 mul_mat for Arc770, add scripts...
2026-04-25 Reese Levineggml-webgpu: support for SSM_SCAN and disable set_rows...
next