Skip to content

server : return language in detect response - #4035

Merged
danbev merged 1 commit into
ggml-org:masterfrom
turicas:fix/detect-language-json
Sep 8, 2026
Merged

danbev merged 1 commit into
ggml-org:masterfrom
turicas:fix/detect-language-json

Conversation

@turicas

@turicas turicas commented Sep 4, 2026

Copy link
Copy Markdown
Contributor

When detect_language=true is used with response_format=json, the server detects the language but returns only an empty text field.

Before this change:

curl -s http://127.0.0.1:8080/inference \
  -F file=@/work/audio.wav \
  -F response_format=json \
  -F detect_language=true \
  -F language=auto
{"text":""}

After this change:

{"text":"","language":"portuguese"}

It uses whisper_full_lang_id(ctx).

Tested with ggml-tiny.bin. Also tested:

cmake --build build --target whisper-server -j2
ctest --output-on-failure -R 'test-whisper-zero-samples|test-whisper-buffer-loader'

Resolves: #3603

@danbev danbev changed the title server : return language in detect response (#3603) server : return language in detect response Sep 8, 2026
@danbev
danbev merged commit 61e6cca into ggml-org:master Sep 8, 2026
46 of 47 checks passed
@turicas

turicas commented Sep 8, 2026

Copy link
Copy Markdown
Contributor Author

Thanks, @danbev!

bygreencn added a commit to bygreencn/whisper.cpp that referenced this pull request Sep 8, 2026
* ggerganov/master:
  whisper : call encoder_begin_callback before language auto-detect (ggml-org#3936)
  whisper : re-seed decoder 0 between calls (ggml-org#4025)
  ci : update close-issue job to not close issues (ggml-org#4045)
  tests : load backends before init when built with GGML_BACKEND_DL (ggml-org#4031)
  server : return language in detect response (ggml-org#4035)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Proper Language detection fixed

2 participants