]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2025-07-04 Georgi Gerganovmetal : disable fast math in all quantize kernels ...
2025-07-04 Georgi Gerganovbatch : add optional for sequential equal split (#14511)
2025-07-04 Georgi Gerganovgraph : prepare for 4D mask (#14515)
2025-07-04 Georgi Gerganovbatch : add n_used count (#14512)
2025-07-04 luyhcsuCANN: Replace aclrtMemsetSync with aclnnInplaceZero...
2025-07-03 Sigbjørn Skjæretggml : implement GEGLU_ERF and GEGLU_QUICK ops (#14445)
2025-07-03 lhezopencl : broadcast for soft_max (#14510)
2025-07-03 Jeff Bolzvulkan: support mixed/deepseekR1 FA head sizes (#14509)
2025-07-03 Johannes Gäßlerggml: backward pass for split swiglu (#14483)
2025-07-03 Nicolò ScipioneFix conditional enabling following arch checks for...
2025-07-03 Xuan-Son Nguyenconvert : correct gemma 3n conversion (#14450)
2025-07-03 Georgi Gerganovkv-cache : use ggml_set_rows (#14285)
2025-07-03 Georgi Gerganovggml : fix FA mask dim 2 and 3 (#14505)
2025-07-03 Georgi Gerganovggml : remove kompute backend (#14501)
2025-07-02 Aman GuptaCUDA: add dynamic shared mem to softmax, refactor gener...
2025-07-02 Sigbjørn Skjæretgguf-py : add support for chat template jinja files...
2025-07-02 compiladellama : initial Mamba-2 support (#9126)
2025-07-02 Georgi Gerganovsync : ggml
2025-07-02 Daniel Beveniusggml : add version function to get lib version (ggml...
2025-07-02 Rotem DanSet RPATH to "@loader_path" / "$ORIGIN" to ensure execu...
2025-07-02 Aman GuptaCUDA: add softmax broadcast (#14475)
2025-07-02 Johannes GäßlerCUDA: broadcasting for FlashAttention mask (#14500)
2025-07-02 Jeff Bolzvulkan: support softmax/FA batch and broadcast (#14449)
2025-07-02 Georgi Gerganovggml : support bcast ggml_soft_max_ext, ggml_flash_attn...
2025-07-02 zhouwgopencl : fix possible buffer overflow in dump_tensor...
2025-07-02 Georgi Gerganovsimple-chat : fix context-exceeded condition (#14494)
2025-07-02 Eric Zhangopencl : skip empty nodes on cgraph compute (#14491)
2025-07-02 lhezopencl : update upscale to support align corners (...
2025-07-02 Sigbjørn Skjæretci : add OpenCL to labeler workflow (#14496)
2025-07-02 Eric Zhanggithub : add OpenCL backend to issue templates (#14492)
2025-07-02 Björn Gansterggml : Callback before abort (#14481)
2025-07-01 Georgi Gerganovci : disable fast-math for Metal GHA CI (#14478)
2025-07-01 Grzegorz GraszaAdd Vulkan images to docker.md (#14472)
2025-07-01 Chenguang LiCANN: update aclnnGroupedMatmulV2 to aclnnGroupedMatmul...
2025-07-01 Jeff Bolzvulkan: Split large mul_mat_id to fit in shared memory...
2025-07-01 Sigbjørn Skjæretadd GELU_ERF (#14455)
2025-07-01 Georgi Gerganovggml : remove trailing whitespace (#0)
2025-07-01 Georgi Gerganovsync : ggml
2025-07-01 Aclyggml-cpu : "align corners" for bilinear upscale/downsca...
2025-07-01 Daniel Beveniusggml-quants : rename best_mad to best_error (ggml/1283)
2025-07-01 lhezopencl : add GEGLU, REGLU, SWIGLU (#14456)
2025-06-30 Aman GuptaAdd Conv2d for CPU (#14388)
2025-06-30 Georgi Gerganovmemory : correctly handle failure in apply() (#14438)
2025-06-30 Georgi Gerganovmetal : disable fast-math for some cpy kernels (#14460)
2025-06-30 Romain Biessyggml-cpu: sycl: Re-enable exp f16 (#14462)
2025-06-30 Diego Devesatest-backend-ops : disable llama test (#14461)
2025-06-30 xiaobing318cmake : Remove redundant include path in CMakeLists...
2025-06-30 Vedran Miletićscripts : make the shell scripts cross-platform (#14341)
2025-06-29 matteoserver : support jinja extra template kwargs (Qwen3...
2025-06-29 Renatserver : fix appearance of the chats list context menu...
2025-06-29 Akarshan BiswasSYCL: disable faulty fp16 exp kernel (#14395)
2025-06-29 Sigbjørn Skjæretggml : fix unmerged GGML_FPxx_TO_FPxx refactoring ...
2025-06-29 Sigbjørn Skjæretggml : implement REGLU/GEGLU/SWIGLU ops (#14158)
2025-06-29 Jeff Bolzvulkan: Add fusion support for RMS_NORM+MUL (#14366)
2025-06-28 Aman GuptaCUDA: add bf16 and f32 support to cublas_mul_mat_batche...
2025-06-28 Jeff Bolzvulkan: handle noncontig in the final case of ggml_vk_g...
2025-06-28 Jeff Bolzvulkan: lock accesses of pinned_memory vector (#14333)
2025-06-28 Weizhao Ouyangmodel : add support for ERNIE 4.5 0.3B model (#14408)
2025-06-28 Xinpeng Doufix async_mode bug (#14432)
2025-06-28 Sigbjørn Skjæretci : fix windows build and release (#14431)
2025-06-28 Jeff Bolzvulkan: Fix GGML_VULKAN_SHADER_DEBUG_INFO (#14427)
2025-06-27 Georgi Gerganovgraph : make llm_graph_context destructor virtual ...
2025-06-27 Georgi Gerganovrecurrent : call balloc split_reset() in init_batch...
2025-06-27 Radoslav Gerganovggml : add ggml_set_rows (#14274)
2025-06-27 Sigbjørn Skjæretconvert : fix broken sentencepiece vocab (#14416)
2025-06-26 Xuan-Son Nguyenmodel : gemma3n text-only (#14400)
2025-06-26 bandoticmake: regen vulkan shaders when shaders-gen sources...
2025-06-26 Sigbjørn Skjæretllama : return mistral-v7-tekken as default template...
2025-06-26 Georgi Gerganovmetal : add special-case mat-vec mul for ne00 == 4... upstream/0.0.5760
2025-06-26 Georgi Gerganovmetal : batch rows copy in a single threadgroup (#14384)
2025-06-26 Aaron Teodocs: update s390x documentation + add faq (#14389)
2025-06-26 R0CKSTARmusa: enable fp16 mma (all) and cublas on qy2 (#13842)
2025-06-25 Aaron Teoggml-cpu: enable IBM NNPA Vector Intrinsics (#14317)
2025-06-25 Sigbjørn Skjæretggml : do not output unprintable characters on GGUF...
2025-06-25 Anton Mitkovsycl: GGML_SYCL_DISABLE_OPT on by default for all Intel...
2025-06-24 lhezopencl: ref count `ggml_backend_opencl_context` and...
2025-06-24 Georgi Gerganovbatch : fix check for empty sequences in memory (#14364)
2025-06-24 Mathieu Baudiercmake : use LLAMA_BUILD_NUMBER when defining LLAMA_INST...
2025-06-24 Nigel Boschserver : move no API key doc to /health (#14352)
2025-06-24 Sigbjørn Skjæretmain : honor --verbose-prompt on interactive prompts...
2025-06-24 Bartowskijinja : Add Mistral-Small-3.2-24B-Instruct-2506.jinja...
2025-06-23 uvosCUDA/HIP: optimize mmv paths taken for HIP devices...
2025-06-23 bandotici: add workflow for relocatable cmake package (#14346)
2025-06-23 Jeff Bolzvulkan: update windows SDK in release.yml (#14344)
2025-06-23 Molly Sophiallama : better rwkv chat template and add missing ...
2025-06-23 Johannes GäßlerCUDA: mul_mat_v support for batch sizes > 1 (#14262)
2025-06-23 Georgi Gerganovkv-cells : fix tracking of seq_pos (#14339)
2025-06-23 Jeff Bolzvulkan: update windows SDK in CI (#14334)
2025-06-22 Ed Addarioquantize : handle user-defined pruning of whole layers...
2025-06-22 Sigbjørn Skjæretgguf-py : fix SpecialVocab parsing when post_processor...
2025-06-22 Ruikai Pengrun : avoid double tokenization (#14327)
2025-06-22 Georgi Gerganovexamples : fix is_first logic for tokenization (#14329)
2025-06-22 uvosHIP: enable vec fattn on RDNA4 (#14323)
2025-06-22 yuisekimtmd : fix Pixtral OOM with large images by capping...
2025-06-22 Sigbjørn Skjæretcommon : use std::string_view now that we target c...
2025-06-22 Aman GuptaCUDA: add mean operation (#14313)
2025-06-21 Sigbjørn Skjæretgguf-py : fix Qwen3-Embedding eos token (#14314)
2025-06-21 Markus TavenrathAdd support for VK_EXT_debug_utils to add labels to...
2025-06-21 Sigbjørn Skjæretgguf-py : fix TemplateProcessing pair when bos/eos...
2025-06-21 Georgi Gerganovmetal : fix thread-safety (#14300)
next