Search: “apple”

kv-cells: look up the n-gram history in the sequence position index (#28040) get_prev_tokens() rebuilt a (seq, pos) -> token hash map on every ubatch by walking all used cells, while llama_kv_cells al

cuda: fuse MoE weighted expert reduction (#25952) cuda : fuse MoE weighted reduction (mul + view + add) The MoE combine tail currently writes weighted expert outputs to global memory before reducing t

arXiv:2608.29540v1 Announce Type: new Abstract: This paper introduces NepScript Genesis, a Neural Architecture Search (NAS) framework for automated Generative Adversarial Network (GAN) discovery, appl

kv-cells: stop the sequence scan once all sequences are seen (#28011) for_each_token_in tested all LLAMA_MAX_SEQ sequences for every used cell, while a cell almost always belongs to one. The scan now

rpc: avoid serializing buffers from other servers (#26500) rpc: avoid serializing buffers from other servers Only include remote buffer pointers when the buffer belongs to the RPC dispatcher receiving

vulkan: top_k radix select for k >= 1024 for Qwen 3.8 Flash Next (#28032) vulkan: add top-k radix sort shader for k >= 1024 add Qwen 3.8 Flash Next top-k tests add top-k qsa fusion clean up code Websi

ggml : add MUL_MAT to the list of ops that may need additional memory (for WebGPU) (#28071) Co-authored-by: Stanisław Szymczyk sszymczy@gmail.com Website: https://llama.app Attestations: https://githu

vulkan: tune mat-vec rows for batched inference on Strix Halo (#27909) vulkan: RDNA3 static mat-vec rows above four columns On RDNA3 above four columns a static 4 rows for all types benches faster tha

spec : fuse the DFlash encoder into the KV cache injection (#27310) dflash : fuse the encoder into the KV injection decode The encoder is a single fc + norm, but running it as a separate llama_encode

sycl : Enhance to get the free memory of Intel GPU (#27968) enhance get mem info by l0 an SYCL API remove debug code, format the code update SYCL.md for GGML_SYCL_GET_MEM_API Website: https://llama.ap

CUDA: extend MOE fusion to specdec, earlier MOE glu fusion and topk-router fusion were restricted to 1 token (#27621) CUDA: extend MOE fusion to specdec, earlier MOE glu fusion and topk-router fusion

ROCm: add radix TOP_K for long rows (#27466) ROCm: add radix TOP_K for long rows Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44178618 macOS/iOS: macOS A

webgpu : avoid crash when offset is not multiple of 4 in WebGPU ggml_backend_tensor_get() implementation (#28045) webgpu : avoid crash when offset is not multiple of 4 in WebGPU ggml_backend_tensor_ge

opencl: tune the quant paths for Intel Xe-LP GPUs to improve its TG and PP performance (#26438) opencl: Q4_K/Q5_K mul_mv N_DST 4->8 on Intel for 2x activation reuse opencl: Q4_K mul_mm 8x8 tile fot In

kv-cache : optimize restoring non-contiguous cells (#27991) kv cache : batch state restore scatter reads per contiguous run When restoring state into non-contiguous destination cells (e.g. a prompt-ca

AVX2: Speed up large batch size prompt processing of IQ models (#27402) Batched gemm for grid IQ quants Style updates and a bit more performance Clean up comments Move code around Vectorize IQ panel d

metal : add concat support for quantized types (#28116) Assisted-by: pi:llama.cpp/DeepSeek-V4-Flash-0731 Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/442

CUDA: XOR swizzle flash attn K,V smem fp16 tiles (#25635) CUDA: XOR swizzle flash attn K,V smem fp16 tiles Signed-off-by: ynankani ynankani@nvidia.com Fix use 64bit generic pointer instead of 32bit sh

metal : add fa-vec tunings for M1 Ultra (#28088) metal : add fa-vec tunings for M1 Ultra metal : move M1 Ultra tunings after M1 Max section metal : remove duplicate blank line Website: https://llama.a

Mac Studio, 96GB unified memory (M3 Ultra). I want the largest usable Qwen3.8-Flash-Next setup, and I'd rather not burn 100GB of bandwidth on the wrong download. Here's my math. Please tell me which p

ggml: allow passing alloc dependencies in graph_optimize (#27301) ggml: allow passing alloc dependencies in graph_optimize add alloc dep tests add TODO about using flat array Website: https://llama.ap

memory : copy Hadamard matrix to k_rot tensor only if it has buffer assigned to prevent crashes during context shift of unquantized K cache (#27967) Co-authored-by: Stanisław Szymczyk sszymczy@gmail.c

metal : fix null-pipeline crash for F16 src1 mul_mat/mul_mat_id (#25648) metal : fail closed on mul_mat shapes with missing F16 kernels metal : abort on nil pipeline in encoder_set_pipeline metal : ad

hexagon: support for device discovery and create sessions on demand (#27785) hex-devices: add support for lazy session allocation and cleanup dev interfaces Co-authored-by: Marco Colombo mcolombo@qti.

vendor : update cpp-httplib to 0.54.0 (#27919) vendor : update cpp-httplib to 0.54.0 vendor : update cpp-httplib to 0.54.0 and 0.54.1 Website: https://llama.app Attestations: https://github.com/ggml-o

ggml : add ggml_backend_op_alloc_size_may_expand, use it in RPC (#27960) some backends (Metal, SYCL, WebGPU) require additional memory for fleeting data for certain ops, which is reflected in their ge

common: rename --tensor-read-lazy to --lazy-mode, add -lzm shorthand (#27969) Rename the --tensor-read-lazy CLI argument to --lazy-mode, to match the internal lazy_mode parameter, and add a -lzm short

dflash: pass missing NVFP4 scales to attention operations (#28000) DFlash2 NVFP4 draft models produced almost no accepted speculative tokens because the Q, K, V, and output projection scales were not

hip : optimize Q2_0 dot-product path for gfx1201 (#26753) hip/gfx1201: optimize q2_0 vec_dot_q2_0_q8_1 with native amdgcn perm Broadened HIP's Q2_0 perm optimization Remove redundant HIP perm availabi

CUDA: use the fast mm_ids_helper path for any n_expert_used (#27978) The optimized path grouped warp lanes by token and required warp_size % n_expert_used == 0, with a single hardcoded exception paddi

llama: improve TENSOR_READ_LAZY handling (#27837) force lazy tensor on cpu if lazy is on llama: improve TENSOR_READ_LAZY handling Website: https://llama.app Attestations: https://github.com/ggml-org/l

vulkan: Change mul_mat_id to pad K rather than N (#27925) The N padding is needed for mul_mat, but not mul_mat_id. For mul_mat_id, we indirect the row index through a shared memory lookup table which

vulkan: combine duplicated fastdiv functions, rename the one optimizing small divs (#27526) vulkan: combine duplicated fastdiv functions, rename the one optimizing small divs remove one more fastdiv W

sycl: make --fit respect --fit-target better (#27629) improve the --fit algorithm to take into account the actual peak required VRAM for a given context size on a SYCL backend. This includes both prop

metal : assert shared memory padding (#27951) metal : assert shared memory padding cont : add ref Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/43867822 m

opencl: use a better matmul path on two Adreno GPU generations (#27640) opencl: default the Adreno xmem F16xF32 GEMM on for X2E kernel_mul_mm_f16_f32_l4_lm is the slowest matmul this backend has on Ad

model: add DSpark support for Nemotron3.5 (#27804) model: add DSpark support for Nemotron3.5 Update src/models/dflash.cpp Co-authored-by: Sigbjørn Skjæret sigbjorn.skjaeret@huggingface.co Co-authored-

tests : run test-save-load-state across all architectures (#27755) tests : run test-save-load-state across all architectures test-save-load-state previously only ran in ctest against a single download

sycl: bind the f16 KV cache in place for the oneDNN SDPA path (#27468) Measured at a live KV length of 34816 (32768 depth plus one 2048 ubatch), on Qwen3.8 27B Q4_K_S: per tensor 4 * 34816 * 256 * 2 B

sycl: use TILE for quantized KV decode on BMG (#26689) Route quantized KV decode to TILE on Xe2 (BMG) only, keep VEC on other archs until validated there. Website: https://llama.app Attestations: http

OpenVINO: Update OV to 2026.3.1, whisper.cpp support, Qwen3.5 on NPU, and new ops (#27843) OpenVINO Backend: Fuse IM2COL + MatMul convolution into OpenVINO convolution ci:ggml-ov: Skip recurrent state

Vulkan: add hoisting support for row IDs and expert count in shaders (#26686) vulkan: add hoisting support for row IDs and expert count in shaders use hoisted row ids in coopmat2 vulkan: address revie

ggml : fix conv_transpose_2d for multiple batches (#26132) ggml : fix conv_transpose_2d for multiple batches ggml_compute_forward_conv_transpose_2d_impl only computed the first batch (ne[3] of the des

vulkan: fix missing view-alias dependencies in ggml_vk_graph_optimize (#27812) vulkan: fix missing view-alias dependencies in ggml_vk_graph_optimize is_src_of doesn't treat two views of one tensor as

bench: add --tensor-read-lazy (#27881) bench: add --tensor-read-lazy rm the alias rename to LLAMA_LAZY_MODE_* Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestation

Hey guys, FastVideo team here. We saw how important speed and quality is for everyone. And we've been working to create our own step distill checkpoints and LORAs for MINIMAX h3. Here's is our v1 rele