]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2025-10-29 Aman GuptaCUDA: Fix bug in topk-moe for gpt-oss (#16821)
2025-10-29 YaelLogicsycl: add RMS_NORM_BACK operation support (#16808)
2025-10-28 YaelGitAccountcuda: add SET operation support (#16804)
2025-10-28 Georgi Gerganovmemory : remove KV cache size padding (#16812)
2025-10-28 Georgi Gerganovllama-bench : clarify benchmarked parts of the computat...
2025-10-28 l3utterflyinitialise buffer.device in ggml_hexagon_session (...
2025-10-28 Sam Malayekembedding: add raw option for --embd-output-format...
2025-10-28 Johannes Gäßlerllama: consistent ctx <-> buf order for KV cache (...
2025-10-28 Aldehir Rojasgrammar : support array references in json schema ...
2025-10-28 Chenguang LiCANN: Improve device ID handling and aclnnArange checks...
2025-10-28 Aman GuptaCUDA: add unused vars to mmvf and mmvq (#16807)
2025-10-28 tamarPalsycl: add SSM_CONV operation support (#16800)
2025-10-27 Yuri Khrustalevchat: Add LFM2 tool handling (#16763)
2025-10-27 Xuan-Son Nguyenmtmd : fix idefics3 preprocessing (#16806)
2025-10-27 Diego Devesallama : disable pipeline parallelism if compute buffer...
2025-10-27 Aclyggml : fix interpolate with align-corners and ne=1...
2025-10-27 Johannes GäßlerHIP: fix AMDGPU_TARGETS, update documentation (#16803)
2025-10-27 Xuan-Son Nguyenmodel : add LightOnOCR-1B model (#16764)
2025-10-27 Johannes Gäßlerllama: fix leaked buffers for mmap + split files (...
2025-10-27 Aman Guptatest-backend-ops: print failed tests at the end (#16785)
2025-10-27 tamarPalsycl: add ROLL operation support (#16665)
2025-10-27 shani-fsycl: add REPEAT_BACK operation support (#16734)
2025-10-27 Aman GuptaCUDA: support for weight clamp in top-k norm (#16702)
2025-10-26 Aclyggml-alloc : make gallocr prefer chunks that allow...
2025-10-26 Sigbjørn Skjæretcuda : use fast copy when src and dst are of different...
2025-10-26 leejetggml: fix cuda kernel launch configuration for k_comput...
2025-10-26 Sigbjørn Skjæretconvert : enable expert group selection for all models...
2025-10-26 Sigbjørn Skjæretgraph : add clamping to ffn_moe_weights_sum to avoid...
2025-10-26 Sigbjørn Skjæretmodel : set res->t_embd in SmallThinker models (#16782)
2025-10-26 amirai21docs : add Jamba to Text-only models list (#16778)
2025-10-26 Aman GuptaCUDA: General GEMV fusion (#16715)
2025-10-26 Gilad S.vulkan: deduplicate Microsoft Direct3D12 devices (...
2025-10-25 Galunidconvert : handle mmproj filename/path properly (#16760)
2025-10-25 Shunta Saitomodel : set res->t_embd in PLaMo2 models (#16766)
2025-10-25 Giuseppe Scrivanovulkan: delete dead code (#16732)
2025-10-25 Jeff Bolzvulkan: Optimize SSM_SCAN (#16645)
2025-10-25 compiladeconvert : avoid dequantizing mxfp4 for GPT-OSS (#16756)
2025-10-24 leejetggml: fix CUDA grid launch condition for large block_nu...
2025-10-24 Aman GuptaCUDA: use CUB for arbitary size argsort (#16754)
2025-10-24 Florian Badiewebui: support q URL parameter (#16728)
2025-10-24 Daniel Beveniusmodel-conversion : add trust_remote_code for orig model...
2025-10-23 compiladeconvert : handle pre-quantized models (#14810)
2025-10-23 Johannes Gäßlerserver: add memory breakdown print (#16740)
2025-10-23 Julien Denizeconvert : Make mistral-common dependency optional ...
2025-10-23 Xuan-Son Nguyenmtmd-cli : allow using --jinja (#16718)
2025-10-23 Prajwal B MehendarkarManually link -lbsd to resolve flock symbol on AIX...
2025-10-23 Aman Guptaggml-cuda: use passed ops instead of hardcoded ops...
2025-10-23 matteoserver : send partial stop string when <EOG> is reached...
2025-10-23 Matthew Michelsycl: use async memory allocation to fix crashes during...
2025-10-22 Max KrasnyanskyAdd experimental ggml-hexagon backend for the Hexagon...
2025-10-22 Diego DevesaRevert "ggml : Leverage the existing GGML_F32_VEC helpe...
2025-10-22 Pascalwebui: introduce OpenAI-compatible model selector in...
2025-10-22 sirus20x6ggml : Leverage the existing GGML_F32_VEC helpers to...
2025-10-22 Aclytests : fix test-thread-safety when compiling with...
2025-10-22 Aman GuptaCUDA: fix bug in topk-moe softmax (#16711)
2025-10-21 Aman GuptaCUDA: topk-moe: add optional parameter for gpt-oss...
2025-10-21 Johannes GäßlerCUDA: better error for FA kernel with 0 occupancy ...
2025-10-21 Aman Guptaggml: add ggml_can_fuse_subgraph (#16662)
2025-10-21 lhezopencl: fix warnings and clean up profiling (#16688)
2025-10-21 Jeff Bolzvulkan: Handle FA with all -inf mask values (#16447)
2025-10-20 YehuditEsycl : add PAD_REFLECT_D1 operator support (#16145)
2025-10-20 Sigbjørn Skjæretmodel : add BailingMoeV2 support (#16063)
2025-10-20 Aleksander... Handle legacy 'context' attachments (#16687)
2025-10-20 Diego Devesaggml-alloc : fix leak when reusing a tensor with a...
2025-10-20 Aleksander... Prevent premature submission on IME input (#16673)
2025-10-20 Aleksander... Import/Export UX improvements (#16619)
2025-10-20 Aleksander... Enable per-conversation loading states to allow having...
2025-10-20 takuya kodamallama-batch: fix build fails with `-Werror=missing...
2025-10-20 Ron Evansreadme: update bindings (#16651)
2025-10-20 safranowithSYCL: Add support for FLOOR,CEIL,ROUND and TRUNC unary...
2025-10-20 takuya kodamallama-context: only warn on pooling_type when user...
2025-10-19 Giuseppe Scrivanomodel : add Granite Hybrid types (#16635)
2025-10-19 Aaron Teoci : fix binaries release failure for s390x (binaries...
2025-10-19 Sigbjørn Skjæretci : avoid manual updates of docs/ops.md (#16663)
2025-10-19 Aaron Teoci: include s390x release binaries (#16648)
2025-10-19 Aman GuptaCODEOWNERS: update for ggml-cuda/mmf (#16660)
2025-10-18 Johannes GäßlerHIP: fix GPU_TARGETS (#16642)
2025-10-18 Jeff Bolzvulkan: Implement topk_moe fused shader, ported from...
2025-10-18 Aman GuptaCUDA: use registers instead of smem in topk-moe (#16647)
2025-10-18 Shawn Guopencl: transposed gemm/gemv moe kernel with mxfp4...
2025-10-17 Johannes Gäßlerllama-model: fix insonsistent ctxs <-> bufs order ...
2025-10-17 Radoslav Gerganovrpc : report actual free memory (#16616)
2025-10-17 Giuseppe Scrivanovulkan: Add State Space Model (SSM) Operations Support...
2025-10-17 muggle-stackggml : fix SpaceMit IME array out-of-bounds in task...
2025-10-17 Pascalwebui: reorganize settings layout (#16607)
2025-10-17 Jeff Bolzvulkan: fix debug build (add_rms_len/data not found...
2025-10-17 Ilia Ilmermetal : add `CONV_TRANSPOSE_2D` (#16542)
2025-10-17 Olivier Chafikgrammar : use int64_t to avoid int overflows in int...
2025-10-17 GittyBursteinSYCL SET operator optimized for F32 tensors (#16350)
2025-10-16 Xuan-Son Nguyenmtmd : support home-cooked Mistral Small Omni (#14928)
2025-10-16 Pascalfix: added a normalization step for MathJax-style ...
2025-10-16 GittyBursteinsycl : add ARANGE operator (#16362)
2025-10-16 Chenguang LiCANN: format code using .clang-format (#15863)
2025-10-16 takasurazeemcommon : Update the docs on -t --threads (#16236)
2025-10-16 takuya kodamaggml-cpu: replace putenv with setenv for const-correctn...
2025-10-16 yael-worksSYCL: Add GGML_OP_MEAN operator support (#16009)
2025-10-15 Aleksei Nikiforovgguf-py : add support for endian conversion of BF16...
2025-10-15 safranowithcpu : add FLOOR, CEIL, ROUND and TRUNC unary operators...
2025-10-15 lhezopencl: add q8_0 mm support (#16469)
2025-10-15 lhezopencl: fix FA for f32 (#16584)
next