Pull Requests ggml-org/whisper.cpp

server : only enable token timestamps when the response needs them

#3990 opened 2026-08-13 22:48 by Kayvan-Zahiri

fix: stop shadowing proper cuda libraries in runtime docker build

#3989 opened 2026-08-13 18:58 by FNGarvin

models : fix n_text_ctx taken from max_length instead of max_target_positions

#3987 opened 2026-08-12 14:38 by axiom-of-choice

whisper : guard null source in buffer loader read callback

#3982 opened 2026-08-09 21:50 by ousamabenyounes

whisper : default-initialize whisper_mel to avoid uninitialized read

#3981 opened 2026-08-09 21:39 by ousamabenyounes

cli : drop invalid UTF-8 bytes from JSON output

#3980 opened 2026-08-09 21:30 by ousamabenyounes

suggestion: add `--continue` flag to `wget` and `wget2`

#3979 opened 2026-08-08 12:25 by OotsutsukiHagoromo

fix: Slow output stage after "saving output to ..." when using --dialize (-di)

#3976 opened 2026-08-07 08:49 by mitslabo

devops : Add main-rocm Dockerfile

#3975 opened 2026-08-06 13:20 by berney

sycl : fix Q8_0/k-quant reorder get_rows (garbage output with GGML_SYCL_ENABLE_OPT)

#3964 opened 2026-08-02 20:28 by xuhancn

ruby : detected non-static command inside ` in extsources.rb...

#3946 opened 2026-07-20 12:40 by anupamme

Fix no_speech_prob calculation

#3945 opened 2026-07-18 19:08 by mculbert

whisper : mask padded K/V in flash-attention encoder and cross-attention

#3941 opened 2026-07-16 09:55 by gooofy

ggml-cpu: gelu optimizqation using simd

#3939 opened 2026-07-09 21:23 by ameya-ms

whisper : run encoder_begin_callback on the language auto-detect path

#3936 opened 2026-07-08 15:29 by javierdejesusda

Fix issue #3269: Account for duration and offset parameters

#3934 opened 2026-07-07 09:49 by mustardpower

Fixes issue: #3924

#3929 opened 2026-07-06 11:10 by michaelsam94

stream : add interactive start/stop mode (-i)

#3926 opened 2026-07-04 12:54 by ailakshya

whisper : reject audio too short for reflective padding

#3925 opened 2026-07-04 10:18 by hellopahe

examples: resolve whisper-cli models from the HuggingFace cache (-hf/-hff)

#3922 opened 2026-07-02 20:39 by djd0723

tests: cover audio buffer decoding

#3911 opened 2026-06-25 22:28 by jamesachurchill

addon.node : support cancelling transcription via AbortSignal

#3909 opened 2026-06-25 08:03 by buxuku

whisper : optional ANEForge encoder backend (Apple Neural Engine)

#3905 opened 2026-06-23 00:24 by sbryngelson

examples : add parakeet-server example

#3904 opened 2026-06-22 14:13 by danbev

ggml : fix ccache launcher problem

#3874 opened 2026-06-10 05:40 by iacore

metal : wind down leftover residency sets at teardown instead of aborting

#3870 opened 2026-06-09 15:48 by AlexCherrypi

whisper : add resumable / streaming transcription API

#3869 opened 2026-06-09 14:46 by AlexCherrypi

fix: gguf: validate tensor dimensions and key names against zero/empty values

#3864 opened 2026-06-07 13:33 by sgInnora

metal: apple ane decoder

#3848 opened 2026-05-31 17:35 by chenqianhe

constrain auto-detect to a provided set of candidate languages

#3822 opened 2026-05-25 00:36 by ilhamsyahids