Skip to content

--diarize random "(speaker ?)" #3546

@slavonnet

Description

@slavonnet

If you send the same audio file with stereo to whisper.cpp , some lines shows "(speaker ?)". If you send it a second time, it's fine. What is the reason and how to fix it?

wav file is stereo from two channels. Chan 1 - Left, Chan 2 = Right.

run args

./build/bin/whisper-cli -vm /root/whisper.cpp/models/ggml-silero-v6.2.0.bin -m models/ggml-large-v3-turbo-q8_0.bin --vad -f /tmp/q-1509-9257299718-20251130-005355-1764453235.353068.wav --language ru --diarize -sns -t 16

whisper build from 19ceec8 commit (main branch)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions