]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2025-05-22 Aaron Teocommon: Include torch package for s390x (#13699)
2025-05-22 Georgi Gerganovserver : pad small embedding batches (#13692)
2025-05-22 Sigbjørn Skjæretgguf-py : correct charsmap parameter typing (#13701)
2025-05-22 Nicolò Scipionesycl : Remove waits from function calls (#13702)
2025-05-22 Ewan CrawfordSYCL: Avoid using with SYCL-Graph for unsupported nodes...
2025-05-21 Henry Linjamäkiopencl: Add support for multiple devices (#12622)
2025-05-21 Henry Linjamäkiopencl: fix couple crashes (#12795)
2025-05-21 Diego Devesareleases : build CPU backend separately (windows) ...
2025-05-21 Georgi Gerganovhparams : support models for which all layers use SWA...
2025-05-21 Georgi Gerganovserver : improve error reporting (#13680)
2025-05-21 antichristHaterconvert : add qwen2vl support for unsloth merges (...
2025-05-21 Sigbjørn Skjæretexamples : switch retrieval to llama_encode (#13685)
2025-05-21 Emmanuel Ferdmangguf-py : display the invalid gguf type (#13687)
2025-05-21 Xuan-Son Nguyenggml : add ggml_gelu_erf() (#13667)
2025-05-21 Robin Davidssonserver : Add the endpoints /api/tags and /api/chat...
2025-05-21 Dorin-Andrei... server : fix first message identification (#13634)
2025-05-21 Georgi Gerganovkv-cache : simplify the interface (#13660)
2025-05-21 Georgi Gerganovmodel : disable SWA for Phi models (#13676)
2025-05-21 R0CKSTARmusa: Upgrade MUSA SDK version to rc4.0.1 and use mudnn...
2025-05-20 Evevulkan: fix warnings (#13626)
2025-05-20 l3utterflymtmd-helper : bug fix to token batching in mtmd (#13650)
2025-05-20 Georgi Gerganovmodel : fix llama4 graph (#13663)
2025-05-20 Georgi Gerganovllama : remove llama_kv_cache_view API + remove depreca...
2025-05-20 Johannes GäßlerCUDA: skip fully masked-out KV in FA vec kernel (#13584)
2025-05-20 Sigbjørn Skjærettests : avoid github urls due to throttling (#13654)
2025-05-20 Svetlozar Georgievsycl: disable reorder for sycl mulmat (#13536)
2025-05-20 0cc4mSet GLM4 blk.*.attn_output.weight, kqv_out-* matmul...
2025-05-20 Georgi Gerganovmetal : fix typo in FA kernel comments (#13651)
2025-05-20 Georgi Gerganovkv-cache : add SWA support (#13194)
2025-05-20 Xinpeng DouCANN: Update CANN model support (#13162)
2025-05-20 Nicolò Scipionesycl : Overcoming workaround for mmap() allocation...
2025-05-19 psocolovskycommon : add load_progress_callback (#13617)
2025-05-19 0cc4mVulkan: Add f32 accumulator support to quantized mul...
2025-05-19 Alberto Cabrera... sycl : backend documentation review (#13544)
2025-05-19 Xuan-Son Nguyenmtmd : add vision support for llama 4 (#13282)
2025-05-19 Alberto Cabrera... ci : upgraded oneAPI version in SYCL workflows and...
2025-05-19 Georgi Gerganovsync : ggml
2025-05-19 Johannes Gäßlermnist: fix segmentation fault (ggml/1227)
2025-05-19 Diego Devesaggml : fix apple OS check in ggml_print_backtrace ...
2025-05-19 Daniel Tangggml : Fix missing backtrace on Linux (ggml/1228)
2025-05-19 Nickfix: check model pointer validity before use (#13631)
2025-05-19 Chenguang LiCANN: Support MOE Model MUL_MAT_ID (#13042)
2025-05-17 Isaac McFadyenserver : added --no-prefill-assistant flag (#13608)
2025-05-17 Gilad S.cmake: use the current build config for vulkan-shaders...
2025-05-17 Georgi Gerganovparallel : add option for non-shared and larger prompts...
2025-05-17 Jeff Bolzvulkan: move common FA code to flash_attn_base.comp...
2025-05-17 Jeff Bolzvulkan: use scalar FA rather than coopmat2 when N=...
2025-05-16 Zllguidance : official v0.7.20 release (no actual change...
2025-05-16 Xuan-Son Nguyenserver : do not return error out of context (with ctx...
2025-05-16 Xuan-Son Nguyenwebui : improve accessibility for visually impaired...
2025-05-16 Xuan-Son Nguyenreadme : add list of dependencies and their license...
2025-05-16 Diego Devesareleases : use arm version of curl for arm releases...
2025-05-16 Georgi Gerganovmetal : add FA-vec kernel for head size 64 (#13583)
2025-05-16 Diego Devesallama : print hint when loading a model when no backend...
2025-05-16 Sigbjørn Skjæretci : add ppc64el to build-linux-cross (#13575)
2025-05-16 Łukasz Ślusarczyksycl : fixed compilation warnings (#13582)
2025-05-15 Olivier Chafikminja: sync (qwen3) (#13573)
2025-05-15 Diego Devesagguf : use ggml log system (#13571)
2025-05-15 Daniel Tanggguf-py : fix disconnect-before-connect in editor-gui...
2025-05-15 Xuan-Son Nguyenconvert : fix conversion for llama 4 (#13567)
2025-05-15 Atharva Dubeysycl: simplify bin_bcast_kernel (#13383)
2025-05-15 Svetlozar Georgievsycl: reordered Q4_K MMVQ (#13109)
2025-05-15 Łukasz Ślusarczyksycl: use oneDNN for matrices multiplication (#12972)
2025-05-15 Diego Devesallama-bench : fix -ot with dl backends (#13563)
2025-05-15 Xuan-Son Nguyenwebui : handle PDF input (as text or image) + convert...
2025-05-15 Piotr Wilkin... server : proper error handling for missing elements...
2025-05-15 Georgi Gerganovbench : handle decode errors (#13548)
2025-05-15 Olivier Chafik`server`: inject date_string in llama 3.x template...
2025-05-14 Georgi Gerganovkv-cache : fix out-of-bounds view during reserve graph...
2025-05-14 Yibo Caiarm64: optimize q6_k_q8_k kernel with i8mm (#13519)
2025-05-14 Olivier Chafik`common`: add partial regex support (#12808)
2025-05-14 Sigbjørn Skjæreteditorconfig : fix trailing whitespace from #13542...
2025-05-14 Gilad S.fix: crash when calling `llama_state_get_size` on a...
2025-05-14 Johannes GäßlerCUDA: fix crash on large batch size for quant. MoE...
2025-05-14 Diego Devesallama : fix quantize with dl backends (#13539)
2025-05-14 Johannes GäßlerCUDA: faster Deepseek FA, add Turing support (#13435)
2025-05-14 Gabe Goodhartfix: Move build_inp_pos to the top of the graph section...
2025-05-14 Georgi Gerganovserver : passthrough the /models endpoint during loadin...
2025-05-14 Xuan-Son Nguyenserver : fix cache_tokens bug with no cache_prompt...
2025-05-14 bandoticmake: simplify vulkan shader test logic (#13263)
2025-05-14 Jeff Bolzvulkan: KHR_coopmat flash attention (#13506)
2025-05-14 Xuan-Son Nguyenwebui : use fflate for more deterministic gzip compress...
2025-05-14 Luca Stefaniwebui: Allow pasting file from clipboard (#13526)
2025-05-14 ddpasadocs: Update link to ggml-org in multimodal.md (#13513)
2025-05-14 Sigbjørn Skjæretscripts : fix compare-llama-bench.py show parameter...
2025-05-14 Jeff Bolzvulkan: workaround FA compile failures on macos (#13517)
2025-05-13 Ed Addarioquantize : improve tensor-type pattern matching (#13033)
2025-05-13 Xuan-Son Nguyenclip : clip.h become private API (⚠️ breaking change...
2025-05-13 Georgi Gerganovmetal : use FA-vec kernel up to batch size 20 (#13496)
2025-05-13 Georgi Gerganovmetal : optimize multi-sequence FA vec kernel (#13493)
2025-05-13 Dan Johanssonggml-cpu: Update KleidiAI to v1.6 and fix include direc...
2025-05-13 Georgi Gerganovbatched-bench : fix pp batch contents (#13492)
2025-05-13 Xuan-Son Nguyenmtmd : remove libllava, remove clip-quantize-cli (...
2025-05-13 Sigbjørn Skjæretscripts : support arbitrary input file formats in compa...
2025-05-13 Gabe Goodhartmodel : Granite MoE shared (#13269)
2025-05-13 Georgi Gerganovsync : ggml
2025-05-12 Diego Devesallama-bench : add defrag-thold, check for invalid range...
2025-05-12 lhezopencl: remove unnecessary assert for `add` (#13257)
2025-05-12 Xuan-Son Nguyenclip : cap max image size 1024 for qwen vl model (...
2025-05-12 Johannes Gäßlerllama/ggml: add LLM training support (#10544)
next