feat: expose no-speech probability in segment #2654
Merged
+15
−3
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
sachaarbonel
changed the title
sachaarbonel
changed the title
ggerganov
approved these changes
Dec 21, 2024
bygreencn
added a commit
to bygreencn/whisper.cpp
that referenced
this pull request
Dec 26, 2024
* ggerganov/master: (49 commits) cli : add --suppress_nst support (ggerganov#2664) cli : add no_speech_thold (ggerganov#2663) cmake : remove hardcoded install rpath server : fix help print ruby : bug fix on callbacks and no_speech_prob (ggerganov#2656) server : add no-speech threshold parameter and functionality (ggerganov#2654) whisper : rename suppress_non_speech_tokens to suppress_nst (ggerganov#2653) server : add option to suppress non-speech tokens (ggerganov#2649) whisper : rename binaries + fix install (ggerganov#2648) ruby : update gem version to v1.3.1 release : v1.7.3 ci : msys enable SDL2 build (ggerganov#2635) ruby : sync ggml (ggerganov#2643) android : try to fix build files : remove old sources sync : ggml talk-llama : sync llama.cpp sync : ggml ggml : update ggml_backend_cpu_device_supports_op (llama/10867) vulkan: bugfixes for small subgroup size systems + llvmpipe test (llama/10809) ...
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds support for exposing no-speech probability at the segment level in Whisper transcriptions, which helps identify potential hallucinations and non-speech segments in the output.
Key Changes
no_speech_probfield to whisper_segment structwhisper_full_get_segment_no_speech_probno_speech_tholdparameter (default: 0.6) via CLI argumentsWhy This Matters