ggerganov/whisper.cpp

Pull Requests Commits

ggml : arm repack fix build (#0)

ggerganov committed 39 days ago

Verified f0c9017a

talk-llama : sync llama.cpp

ggerganov committed 40 days ago

Verified 179d8b1c

ggerganov committed 40 days ago

Verified 48cdc06e

whisper : adjust to ggml changes (#0)

ggerganov committed 40 days ago

Verified 72714d16

cmake : set `CMAKE_RUNTIME_OUTPUT_DIRECTORY` for non standalone build (ggml/1394)

HerrCai0907 committed 40 days ago

Verified 324dd21d

ggml-alloc : fix reuse-parent logic for misaligned sizes (llama/17884)

ggerganov committed 40 days ago

Verified 1da1a686

ggml-hexagon: fix `rope` failure at `test-backend-ops` (llama/17565)

chraac committed 40 days ago

Verified 0c88de5c

Fix race conditions in threadpool when dealing with dynamic/frequent n_threads changes (llama/17748)

max-krasnyansky committed 40 days ago

Verified a2886fba

ggml : remove GGML_KQ_MASK_PAD constant (llama/17910)

ggerganov committed 40 days ago

Verified cd9b8c6d

cuda : add missing support check for xielu (llama/17895)

CISC committed 40 days ago

Verified ca8ea18d

CUDA: fix unpadded strides in MMA FA kernel (llama/17891)

JohannesGaessler committed 40 days ago

Verified ea182913

fix softmax for iGPU (llama/17838)

NeoZhangJianyu committed 40 days ago

Verified c10b4f9a

metal: SSM kernel improvements (llama/17876)

gabe-l-hart committed 40 days ago

Verified 307dc525

Add DIAG for CUDA (llama/17873)

pwilkin committed 40 days ago

Verified 2817582b

ggml : Provide macos-specific backtrace printing to avoid terminal death (llama/17869)

gabe-l-hart committed 40 days ago

Verified 41bbc034

metal : print node names for debugging (llama/17882)

ggerganov committed 40 days ago

Verified b6ae0b29

ggml : allow fill node alloc inplace (llama/17870)

CISC committed 40 days ago

Verified ba463fb5

CANN: add support for partial RoPE and Vision mode (llama/17543)

noemotiovon committed 40 days ago

Verified 79d86a5c

CUDA: fix FP16 overflow in tile FA kernel (llama/17875)

JohannesGaessler committed 40 days ago

Verified bef1f5a5

cuda : add FILL op support (llama/17851)

JayZenith committed 40 days ago

Verified 821c2071

cuda: optimize SOLVE_TRI using registers and FMAF (llama/17703)

wsbagnsv1 committed 40 days ago

Verified e1562e85

ggml-cpu: add ggml_thread_cpu_relax with Zihintpause support (llama/17784)

ixgbe committed 40 days ago

Verified c8d0ee2f

Vulkan: improve mul_mat_vec_iq1_m (llama/16907)

lovedheart committed 40 days ago

Verified d6d44fac

sycl: add missing BF16 conversion support for Intel oneAPI (llama/17780)

yingying0906 committed 40 days ago

Verified 447ef863

vulkan: perf_logger improvements (llama/17672)

jeffbolznv committed 40 days ago

Verified 898f876f

ggml-zendnn : add ZenDNN backend for AMD CPUs (llama/17690)

z-vishal committed 40 days ago

Verified ebff8f9d

ggml : add circular tiling support to pad, for Vulkan, CUDA, and CPU (used for making seamless textures) (llama/16985)

Phylliida committed 40 days ago

Verified c5e18070

HIP: fix RDNA3 FP16/BF16 matrix multiplication (llama/17817)

JohannesGaessler committed 40 days ago

Verified 94be7191

ggml : improve error handling for search path existence checks (llama/17653)

flyinskyin2013 committed 40 days ago

Verified b67e3abd

vulkan: Use one row per workgroup for f32 mmv (llama/17711)

jeffbolznv committed 40 days ago

Verified c66c71e9

Older