ggerganov/whisper.cpp

Pull Requests Commits

Copilot committed 43 days ago

7d79ef9f

sync : llama.cpp

ggerganov committed 44 days ago

b12abefa

ggerganov committed 44 days ago

0e5deca8

metal : support I32 -> I32 copy (llama/17317)

ggerganov committed 44 days ago

66156735

metal : faster argsort (llama/17315)

ggerganov committed 44 days ago

74bb8a8b

metal : add cumsum (llama/17305)

ggerganov committed 44 days ago

57c0e6f8

CANN: Use smart pointers to manage ACL objects (llama/17238)

hipudding committed 44 days ago

d3f54874

vulkan: add LOG operation support for F32 and F16 (llama/17183)

zayac committed 44 days ago

9d95d9a1

vulkan: fix MMQ quantize_y condition (llama/17301)

0cc4m committed 44 days ago

f571655e

metal : remove obosolete asserts (llama/17295)

ggerganov committed 44 days ago

9549cc10

opencl: fix rms_norm_mul (llama/17250)

lhez committed 44 days ago

a75525ca

opencl: add kernel to handle mat mul in attention to improve encoding speed (llama/17181)

shaofeiqi committed 44 days ago

c78845bf

sycl : unify unary kernels with a generic implementation and enable wide operator support (llama/17213)

shani-f committed 44 days ago

1fd63da9

vulkan: Fuse mul_mat_id+add_id+mul and mul_mat+add+add. (llama/17287)

jeffbolznv committed 44 days ago

ea3ebd8b

vulkan: Replace 16-bit unpack8 calls to work around legacy Windows AMD driver bug (llama/17285)

0cc4m committed 44 days ago

7caea544

vulkan: implement ABS and NEG (llama/17245)

giuseppe committed 44 days ago

4c4e663d

vulkan: Use ggml_vk_tensor_subbuffer in mul_mat_vec(id) paths (llama/17244)

jeffbolznv committed 44 days ago

e1846fc5

vulkan: skip all-negative-inf blocks in FA (llama/17186)

jeffbolznv committed 44 days ago

9614a563

vulkan: change graph_compute to be async and enable get_tensor_async (llama/17158)

jeffbolznv committed 44 days ago

37d4bba1

metal : support argsort for ne00 > 1024 (llama/17247)

ggerganov committed 44 days ago

523a6c27

metal : make the FA extra sizes consistent (llama/17143)

ggerganov committed 44 days ago

b4d7df3b

ggml-cpu: handle 3d tensors in repack mat_mul (llama/17241)

Alcpz committed 44 days ago

a81fbfc7

ggml : add ops SOFTPLUS, EXPM1, TRI, SOLVE_TRI, CUMSUM (llama/17063)

pwilkin committed 44 days ago

3e684f26

vulkan: remove shell call from vulkan-shaders-gen tool, revert file check (llama/17219)

0cc4m committed 44 days ago

e8e0004f

sched : fix reserve ignoring user tensor assignments (llama/17232)

slaren committed 44 days ago

210f0f86

ggml-cpu : add RISC-V vector intrinsic support for silu and cvar operations (llama/17227)

ixgbe committed 44 days ago

91fa5b5c

metal: accelerated conv2d (llama/17175)

bghira committed 44 days ago

265d326f

Revert "ggml-cpu: handle 3d tensors in repack mat_mul (llama/17030)" (llama/17233)

ggerganov committed 44 days ago

6a1d830d

ggml-cpu : use template for argsort (llama/17222)

slaren committed 44 days ago

6a91780c

CANN: Add cross_entropy_loss op support (llama/16886)

TecJesh committed 44 days ago

726912d1

Older