All pull requests

Pull requests

• UPSTREAM PR #30150: ggml-cpu : thread fixed-width concat over all rows, constant element …

  #1367·loci-dev opened Oct 9, 2026·

  ·
  1

• UPSTREAM PR #29835: vulkan: Add workaround for device free memory > total memory on Intel Windows

  #1366·loci-dev opened Oct 9, 2026·

  ·
  1

• UPSTREAM PR #29836: ggml-zdnn: introduce Q8_0 mat mul

  #1365·loci-dev opened Oct 8, 2026·

  ·
  12

• UPSTREAM PR #29200: cpu: add F16 SWIGLU_OAI reference implementation and validation

  #1364·loci-dev opened Oct 8, 2026·

  ·
  23

• Fix delayed AllReduce on Gemma-4 MoE

  #1363·msmilevski opened Oct 2, 2026·

  ·

• UPSTREAM PR #22102: fix: GLM-DSA crash in llama-tokenize when using vocab_only

  #1360·loci-dev opened Apr 19, 2026·

  ·
  1

• UPSTREAM PR #22101: mtmd: add granite-speech support (ibm-granite/granite-4.0-1b-speech)

  #1359·loci-dev opened Apr 19, 2026·

  ·
  1

• UPSTREAM PR #22070: ggml-cuda: gate native ue4m3 conversion to sm_90+

  #1358·loci-dev opened Apr 18, 2026·

  ·
  1

• UPSTREAM PR #22066: sycl: Battlemage (BMG) optimizations — AOT, Q5_K reorder, PAD stride fix, new ops, oneMKL routing

  #1357·loci-dev opened Apr 18, 2026·

  ·
  1

• UPSTREAM PR #22023: mtmd: add missing struct tag

  #1356·loci-dev opened Apr 17, 2026·

  ·

• UPSTREAM PR #21647: ci: add android arm64 build and release

  #1355·loci-dev opened Apr 17, 2026·

  ·

• UPSTREAM PR #21985: Feat/modelscope support

  #1354·loci-dev opened Apr 16, 2026·

  ·
  2

• UPSTREAM PR #21071: hexagon: optimize HMX matmul operations

  #1351·loci-dev opened Apr 15, 2026·

  ·
  1

• UPSTREAM PR #21652: Prevent the sum of the dequantized activation in q8_1 from overflowing

  #1350·loci-dev opened Apr 15, 2026·

  ·
  1

• UPSTREAM PR #21870: common: skip reasoning budget sampler when no budget is requested

  #1349·loci-dev opened Apr 14, 2026·

  ·
  1

• UPSTREAM PR #21835: Expose build_info in router mode

  #1348·loci-dev opened Apr 13, 2026·

  ·
  1

• UPSTREAM PR #21821: llama : add --hugepages for HugeTLB-backed weight loading (Linux)

  #1347·loci-dev opened Apr 13, 2026·

  ·
  1

• UPSTREAM PR #21554: hexagon: optimization for HMX mat_mul

  #1346·loci-dev opened Apr 12, 2026·

  ·
  1

• UPSTREAM PR #21787: vulkan: fix output corruption on GCN 2.0/3.0 (Vulkan 1.2)

  #1345·loci-dev opened Apr 12, 2026·

  ·
  1

• UPSTREAM PR #21753: vulkan: Support asymmetric FA in coopmat2 path

  #1344·loci-dev opened Apr 11, 2026·

  ·
  1

• UPSTREAM PR #21629: cuda: Q1_0 initial backend

  #1343·loci-dev opened Apr 11, 2026·

  ·
  1

• UPSTREAM PR #21344: gfx1151 nwarps, tile sizing to curb VGPR pressure

  #1342·loci-dev opened Apr 10, 2026·

  ·
  1

• UPSTREAM PR #21431: vulkan: Tweak Xe2 warptile configuration

  #1341·loci-dev opened Apr 10, 2026·

  ·
  1

• UPSTREAM PR #21597: SYCL: fix multi-GPU system RAM exhaustion by using Level Zero allocations

  #1340·loci-dev opened Apr 8, 2026·7 tasks done·

  ·
  1

• UPSTREAM PR #21573: autoparser: fix MiniMax handling

  #1339·loci-dev opened Apr 8, 2026·

  ·
  1