Commit History

Автор SHA1 Съобщение Дата
  Shawn Gu 81387858f1 opencl: transposed gemm/gemv moe kernel with mxfp4,f32 (#16602) преди 3 месеца
  Johannes Gäßler 66b0dbcb2d llama-model: fix insonsistent ctxs <-> bufs order (#16581) преди 3 месеца
  Radoslav Gerganov 41386cf365 rpc : report actual free memory (#16616) преди 3 месеца
  Giuseppe Scrivano 3d4e86bbeb vulkan: Add State Space Model (SSM) Operations Support (#16463) преди 3 месеца
  muggle-stack 342c728d03 ggml : fix SpaceMit IME array out-of-bounds in task assignment (#16629) преди 3 месеца
  Pascal ababae7e1e webui: reorganize settings layout (#16607) преди 3 месеца
  Jeff Bolz b19491599d vulkan: fix debug build (add_rms_len/data not found) (#16624) преди 3 месеца
  Ilia Ilmer 9ad4f1931e metal : add `CONV_TRANSPOSE_2D` (#16542) преди 3 месеца
  Olivier Chafik 79967ec596 grammar : use int64_t to avoid int overflows in int schema to grammar conversion logic (#16626) преди 3 месеца
  GittyBurstein ceff6bb253 SYCL SET operator optimized for F32 tensors (#16350) преди 3 месеца
  Xuan-Son Nguyen 1bb4f43380 mtmd : support home-cooked Mistral Small Omni (#14928) преди 3 месеца
  Pascal 683fa6ba4e fix: added a normalization step for MathJax-style \[\] and \(\) delimiters (#16599) преди 3 месеца
  GittyBurstein b22572e97d sycl : add ARANGE operator (#16362) преди 3 месеца
  Chenguang Li 7a50cf388a CANN: format code using .clang-format (#15863) преди 3 месеца
  takasurazeem 6f5d924637 common : Update the docs on -t --threads (#16236) преди 3 месеца
  takuya kodama adc9b60f19 ggml-cpu: replace putenv with setenv for const-correctness (#16573) преди 3 месеца
  yael-works ee50ee1ead SYCL: Add GGML_OP_MEAN operator support (#16009) преди 3 месеца
  Aleksei Nikiforov 7adc79c032 gguf-py : add support for endian conversion of BF16 data (#16594) преди 3 месеца
  safranowith 466c1911ab cpu : add FLOOR, CEIL, ROUND and TRUNC unary operators (#16083) преди 3 месеца
  lhez 0cb7a0683b opencl: add q8_0 mm support (#16469) преди 3 месеца
  lhez d93f8439b0 opencl: fix FA for f32 (#16584) преди 3 месеца
  Aleksander Grygier f9fb33f263 Add server-driven parameter defaults and syncing (#16515) преди 3 месеца
  Sam/Samuel f4ce81c45e metal: optimise `GGML_OP_SUM` (#16559) преди 3 месеца
  Georgi Gerganov 17304cbcc1 server : fix img token logs (#16595) преди 3 месеца
  Xuan-Son Nguyen 3e3cb19f64 llama-quant: add support for mmproj (#16592) преди 3 месеца
  Julius Tischbein 5acd455460 CUDA: Changing the CUDA scheduling strategy to spin (#16585) преди 3 месеца
  Georgi Gerganov 554fd578a5 server : fix mtmd checkpoints (#16591) преди 3 месеца
  Georgi Gerganov fa882fd2b1 metal : avoid using Metal's gpuAddress property (#16576) преди 3 месеца
  SavicStefan ffa059034c vulkan: Add ACC_TYPE_VEC2 implementation (#16203) преди 3 месеца
  Aman Gupta 120bf7046d CUDA + openCL: fix bug in accessing rms_norm->src while doing fusion (#16577) преди 3 месеца