]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/shortlog
pkg/ggml/sources/whisper.cpp
2024-06-16 Georgi Gerganovggml : restore ggml_rope_xpos_inplace (ggml/0)
2024-06-16 Masaya, Katoggml: aarch64: SVE kernels for q8_0_q8_0, q4_0_q8_0...
2024-06-16 Georgi Gerganovggml : silence UB sanitizer error during iq2_xxs quanti...
2024-06-16 Georgi Gerganovggml : remove ggml_flash_attn and ggml_flash_ff (llama...
2024-06-16 Georgi Gerganovggml : drop support for QK_K=64 (llama/7473)
2024-06-16 0cc4mUpdate vulkan rope implementation to support frequency...
2024-06-16 Johannes GäßlerCUDA: fix FA out-of-bounds reads (llama/7479)
2024-06-16 Johannes GäßlerCUDA: fix FA out-of-bounds writes (llama/7465)
2024-06-16 Georgi Gerganovcuda : fix compile warning (llama/7454)
2024-06-16 Johannes GäßlerCUDA: remove incorrect precision check (llama/7454)
2024-06-16 Georgi Gerganovcuda : fix rope + add tests (llama/7452)
2024-06-16 liuwei-gitllama : add phi3 128K model support (llama/7225)
2024-06-16 Georgi Gerganovmetal : handle F16 inf values, fix FA partial offload...
2024-06-16 Johannes GäßlerCUDA: fix unused warning in mmq.cu (llama/7442)
2024-06-16 Johannes GäßlerCUDA: deduplicate mmq code (llama/7397)
2024-06-16 Radoslav Gerganovrpc : track allocated buffers (llama/7411)
2024-06-16 AidanBeltonSUpdate SYCL upscale operation (llama/7321)
2024-06-16 Herman Semenovggml-opencl, llama: using reserve() if count already...
2024-06-16 junchao-loongsonggml : add loongarch lsx and lasx support (llama/6454)
2024-06-16 Srihari-mcwAdd provisions for windows support for BF16 code includ...
2024-06-16 0cc4mVulkan Embedding Fix (llama/7360)
2024-06-16 slarenggml : fix another case of quants nans (llama/7387)
2024-06-16 Johannes Gäßlerggml: implement quantized KV cache for FA (llama/7372)
2024-06-16 slarencuda : clear error after buffer allocation failure...
2024-06-16 fraxy-vCapture CUDA logging output (llama/7298)
2024-06-16 Georgi Gerganovandroid : use "ci-android" branch for CI (llama/7341)
2024-06-16 Johannes GäßlerCUDA: deduplicate FlashAttention code (llama/7352)
2024-06-16 Engininja2cuda : add half2 __shfl_xor() for ROCm 5.5 (llama/7263)
2024-06-16 0cc4mUpdate and fix Vulkan soft_max and argsort implementati...
2024-06-16 slarenggml : fix quants nans when all the group weights are...
2024-06-16 Johannes GäßlerCUDA: faster large batch FA without tensor cores (llama...
2024-06-16 Radoslav Gerganovrpc : set SO_REUSEADDR for the server socket (llama...
2024-06-16 Herman Semenovggml-quants, llama : removed excess checks (llama/7274)
2024-06-16 Justine Tunneyggml : rewrite silu and softmax for cpu (llama/7154)
2024-06-16 Radoslav Gerganovrpc : add command line arg for specifying backend memory
2024-06-16 Max KrasnyanskyAdd support for properly optimized Windows ARM64 builds...
2024-06-16 kunnisggml : use dynamic thread scheduling for matrix multipl...
2024-06-16 agray3Avoid unnecessarily disabling CUDA graphs (llama/7302)
2024-06-16 slarenggml : tag ggml_tensor::backend as deprecated (llama...
2024-06-16 AidanBeltonSAdd missing " (llama/7303)
2024-06-16 John Balisggml : add `ggml_upscale_ext` (ggml/814)
2024-06-16 Georgi Gerganovscripts : update sync
2024-06-13 Borislav Stanimirovwhisper : use ggml-cuda in mel calc, set appropriate...
2024-06-11 Georgi Gerganovcuda : fix HIPBLAS build (#2234)
2024-06-11 Georgi Gerganovcuda : fix bounds check for src0 rows in MMVQ kernel...
2024-06-11 Georgi Gerganovci : fix CUDA builds (#2232)
2024-06-10 Borislav Stanimirovwhisper : auto-grow working areas for mel_calc_cuda...
2024-06-10 Georgi Gerganovwhisper : free whisper_mel instances (#2220)
2024-06-06 Georgi Gerganovwhisper : whisper_state/backend fixes (#2217)
2024-06-06 Borislav Stanimirovwhisper : calculate mel spectrogram directly into a...
2024-06-04 Borislav Stanimirovwhisper : add CUDA-specific computation mel spectrogram...
2024-05-31 Borislav Stanimirovwhisper : remove `speed_up` and `phase_vocoder*` functi...
2024-05-30 Martin Delillereadme : add conan badge (#2196)
2024-05-30 Carlos Zoidoreadme : add install instructions for Conan (#2189)
2024-05-29 Borislav Stanimirovwhisper: use global cache for sin/cos vals and Hann...
2024-05-27 Georgi Gerganovrelease : v1.6.2
2024-05-27 Georgi GerganovRevert "whisper : remove extra backend instance (huh...
2024-05-25 Daniel Valdiviaserver : fix typo (#2181)
2024-05-22 Toddruby : update bindings (#2154)
2024-05-21 Georgi Gerganovrelease : v1.6.1
2024-05-21 William Tambelliniexamples : add support for decoding input with ffmpeg...
2024-05-20 Pedro Probstnode : add flash_attn param (#2170)
2024-05-19 Tamotsu Takahashici: Update build.yml to suppress warnings about node...
2024-05-15 Georgi Gerganovrelease : v1.6.0
2024-05-15 Georgi Gerganovwhisper : use flash attention (#2152)
2024-05-14 petterreinholdtsentalk-llama : reject runs without required arguments...
2024-05-14 Georgi Gerganovsync : ggml
2024-05-14 Georgi Gerganovmetal : support FA without mask + add asserts (llama...
2024-05-14 Radoslav Gerganovggml : add RPC backend (llama/6829)
2024-05-14 Neo Zhangrm wait() (llama/7233)
2024-05-14 Johannes GäßlerCUDA: add FP32 FlashAttention vector kernel (llama...
2024-05-14 Georgi Gerganovscripts : sync ggml-rpc
2024-05-14 thewh1teaglewhisper : fix model path encoding in windows (#2086)
2024-05-13 Georgi Gerganovserver : return utf-8 (#2138)
2024-05-13 Pedro Probstnode : add audio_ctx and audio buffer params (#2123)
2024-05-13 aldorofcmake : fix HIP/ROCm build (#2102)
2024-05-13 valVknode : add additional params (#2000)
2024-05-13 Mark Karpelèsjs : remove un-needed request header from fetchRemote...
2024-05-13 Georgi Gerganovcmake : fix metal embed sources path (#2110)
2024-05-13 Daniel Ziegenbergmain : dont print timings with --no-prints (#2108)
2024-05-13 Daniel Ziegenbergmain : add options for temperature control (#2088)
2024-05-13 Georgi Gerganovwhisper : switch back to F32 mask (#0)
2024-05-13 zhangjixiongwhisper.android : update example, add field to print...
2024-05-13 Xingchen Song... cmake : fix json INTERFACE library (#2069)
2024-05-13 mashizoramain : fix double quote escaping in csv output (#2090)
2024-05-13 Georgi Gerganovmetal : tune soft_max number of threads (#0)
2024-05-13 Georgi Gerganovwhisper : remove old flash attn code (#0)
2024-05-13 Georgi Gerganovggml : try fix ppc64 (#0)
2024-05-13 Georgi Gerganovggml : remove oboslete alibi code (skipme) (#0)
2024-05-13 Georgi Gerganovtalk-llama : sync llama.cpp
2024-05-13 Georgi Gerganovsync : ggml
2024-05-13 Hong Bo PENGggml : optimize for ppc64le using VSX intrinsics (ggml...
2024-05-13 Georgi Gerganovmetal : fix indent (ggml/0)
2024-05-13 Georgi Gerganovggml : restore sigmoid decl order (ggml/0)
2024-05-13 Georgi Gerganovggml : resolve merge (ggml/0)
2024-05-13 Georgi Gerganovggml : full ALiBi support (llama/7192)
2024-05-13 Georgi Gerganovmetal : fix flash attention kernel requirements (llama...
2024-05-13 Ouadie EL FAROUKIMinor arithmetic improvement to mmvq wrapper kernel...
2024-05-13 0cc4mVulkan Bugfixes and Improvements (llama/7084)
2024-05-13 Johannes GäßlerCUDA: generalize FP16 fattn vec kernel (llama/7061)
next