Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Silero VAD for cleaning the dataset from silence #1166

Open
wants to merge 94 commits into
base: master
Choose a base branch
from
Open

Silero VAD for cleaning the dataset from silence #1166

wants to merge 94 commits into from

Commits on Sep 23, 2023

  1. initialise the script for activity detection

    @rilshok
    rilshok committed Sep 23, 2023
    Configuration menu
    Copy the full SHA
    8636a80 View commit details
    Browse the repository at this point in the history

  2. init lhotse.workflows.activity_distillation module

    @rilshok
    rilshok committed Sep 23, 2023
    Configuration menu
    Copy the full SHA
    6c50efe View commit details
    Browse the repository at this point in the history

  3. add the Silero VAD model wrapper

    @rilshok
    rilshok committed Sep 23, 2023
    Configuration menu
    Copy the full SHA
    9fb33f2 View commit details
    Browse the repository at this point in the history

  4. inherit SileroVAD from ActivityDetector

    @rilshok
    rilshok committed Sep 23, 2023
    Configuration menu
    Copy the full SHA
    0419467 View commit details
    Browse the repository at this point in the history

Commits on Sep 24, 2023

  1. pass parameters to the model explicitly

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    bbc5d09 View commit details
    Browse the repository at this point in the history

  2. process each channel and return the supervision

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    7011d06 View commit details
    Browse the repository at this point in the history

  3. parallel processing by activity detector

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    ae29304 View commit details
    Browse the repository at this point in the history

  4. number the segments found

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    c686432 View commit details
    Browse the repository at this point in the history

  5. make abstract processing of an individual track

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    8ad86be View commit details
    Browse the repository at this point in the history

  6. rename module and workflow to activity_detection

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    c42fafe View commit details
    Browse the repository at this point in the history

  7. standardise detectors by sampling rate

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    e2db456 View commit details
    Browse the repository at this point in the history

  8. rename silero vad models

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    9117d98 View commit details
    Browse the repository at this point in the history

  9. implement a script for supervisory with silero-vad

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    bd279ae View commit details
    Browse the repository at this point in the history

  10. handle exceptions and user input

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    9279bd9 View commit details
    Browse the repository at this point in the history

  11. allow the path to the output dir

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    3cf621e View commit details
    Browse the repository at this point in the history

  12. reset the cached state of the model if necessary

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    1a2acdc View commit details
    Browse the repository at this point in the history

  13. add docs for activity_detection module

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    3638360 View commit details
    Browse the repository at this point in the history

  14. fix if dir does not exist

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    1d7cf8c View commit details
    Browse the repository at this point in the history

  15. fix cuda issue

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    2d9e258 View commit details
    Browse the repository at this point in the history

  16. add RecordingSet python example

    @rilshok
    rilshok committed Sep 24, 2023
    Configuration menu
    Copy the full SHA
    7ce7f32 View commit details
    Browse the repository at this point in the history

Commits on Sep 25, 2023

  1. add base test for silero vad workflow

    @rilshok
    rilshok committed Sep 25, 2023
    Configuration menu
    Copy the full SHA
    b335187 View commit details
    Browse the repository at this point in the history

  2. add test for silero vad in parallel

    @rilshok
    rilshok committed Sep 25, 2023
    Configuration menu
    Copy the full SHA
    819e020 View commit details
    Browse the repository at this point in the history

Commits on Sep 26, 2023

  1. change detector name

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    7bd2c69 View commit details
    Browse the repository at this point in the history

  2. replace the chore option with force_download

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    729f279 View commit details
    Browse the repository at this point in the history

  3. improve user experience

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    1d33460 View commit details
    Browse the repository at this point in the history

  4. add simple test for activity_detection workflow

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    80163fe View commit details
    Browse the repository at this point in the history

  5. clarify the need to use --force_download

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    1bc6d6e View commit details
    Browse the repository at this point in the history

  6. rm slash

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    c04558e View commit details
    Browse the repository at this point in the history

  7. trust the repository since torch>=1.12

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    41a9085 View commit details
    Browse the repository at this point in the history

  8. skip tests if torch version <1.12

    @rilshok
    rilshok committed Sep 26, 2023
    Configuration menu
    Copy the full SHA
    5544c25 View commit details
    Browse the repository at this point in the history

Commits on Sep 27, 2023

  1. Merge branch 'master' into feature/distill-cutset

    @pzelasko
    pzelasko authored Sep 27, 2023
    Configuration menu
    Copy the full SHA
    4a342d2 View commit details
    Browse the repository at this point in the history

  2. exclude non-coverage eligible code

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    03a17ba View commit details
    Browse the repository at this point in the history

  3. Merge branch 'feature/distill-cutset' of github.com:rilshok/lhotse in… 

    …to feature/distill-cutset
    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    0bb19d9 View commit details
    Browse the repository at this point in the history

  4. init speach only script

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    0d5db2a View commit details
    Browse the repository at this point in the history

  5. prepare Recording for speech analysis

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    29443d6 View commit details
    Browse the repository at this point in the history

  6. add name and sampling_rate props

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    1edc8b6 View commit details
    Browse the repository at this point in the history

  7. analyze audio and select speech fragments

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    4ed0669 View commit details
    Browse the repository at this point in the history

  8. build an activity tree

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    ef6466b View commit details
    Browse the repository at this point in the history

  9. rename as to_mono

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    aec7c7c View commit details
    Browse the repository at this point in the history

  10. process CutSet in sequential mode

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    4e7f5b9 View commit details
    Browse the repository at this point in the history

  11. fix cut type definition error. +verbosity

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    3ec18f9 View commit details
    Browse the repository at this point in the history

  12. change the behaviour of the force_download option

    @rilshok
    rilshok committed Sep 27, 2023
    Configuration menu
    Copy the full SHA
    ed71b42 View commit details
    Browse the repository at this point in the history

Commits on Sep 28, 2023

  1. convert activity intervals into a trimming tree

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    cc8a19e View commit details
    Browse the repository at this point in the history

  2. trimm numpy array by activity tree

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    e5f8726 View commit details
    Browse the repository at this point in the history

  3. save audio to RAM and generate a new Recording

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    c9eecb6 View commit details
    Browse the repository at this point in the history

  4. transform supervision according to trimming_tree

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    a1ed184 View commit details
    Browse the repository at this point in the history

  5. find the activity, trim and redefine the Cut

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    11e85e0 View commit details
    Browse the repository at this point in the history

  6. trim the AlignmentItem s

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    1bedb5f View commit details
    Browse the repository at this point in the history

  7. save cut only after processing

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    2ce97de View commit details
    Browse the repository at this point in the history

  8. export public names, make system functions private

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    e155706 View commit details
    Browse the repository at this point in the history

  9. move recording to disc by torchaudio

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    479a66d View commit details
    Browse the repository at this point in the history

  10. protect supervision outside of audio recording

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    3e5dec7 View commit details
    Browse the repository at this point in the history

  11. Merge branch 'feature/distill-cutset' into feature/speech-only

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    ebf86ee View commit details
    Browse the repository at this point in the history

  12. Merge branch 'master' into feature/distill-cutset

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    85bc551 View commit details
    Browse the repository at this point in the history

  13. Merge branch 'feature/distill-cutset' into feature/speech-only

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    d422c92 View commit details
    Browse the repository at this point in the history

  14. initialisation of the workflow leaving only speech

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    bdbd5a8 View commit details
    Browse the repository at this point in the history

  15. prepare the input data and run the calculations

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    d2aa228 View commit details
    Browse the repository at this point in the history

  16. save results to output_dir

    @rilshok
    rilshok committed Sep 28, 2023
    Configuration menu
    Copy the full SHA
    c2dcb76 View commit details
    Browse the repository at this point in the history

Commits on Sep 29, 2023

  1. If it does not save anywhere, store it in memory

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    50a7907 View commit details
    Browse the repository at this point in the history

  2. add an iterable multiprocessor executor

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    5a9f5f6 View commit details
    Browse the repository at this point in the history

  3. pass arguments as keywords to ProcessWorker

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    1f6689f View commit details
    Browse the repository at this point in the history

  4. use multiprocessing to speed up processing

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    c0e09ab View commit details
    Browse the repository at this point in the history

  5. Use Processor instead of ActivityDetectionProcessor

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    2a45985 View commit details
    Browse the repository at this point in the history

  6. remove multiprocessing in detect_activity

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    bae565b View commit details
    Browse the repository at this point in the history

  7. save the file inside detect_activity

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    09a7775 View commit details
    Browse the repository at this point in the history

  8. get rid ActivityDetectionProcessor class

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    e7e0278 View commit details
    Browse the repository at this point in the history

  9. move Processor to workflows.backend

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    c9f4d84 View commit details
    Browse the repository at this point in the history

  10. make ActivityDetector's forward abstract __call__

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    43dca35 View commit details
    Browse the repository at this point in the history

  11. rename the workflow to trim_inactivity

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    13ca2e9 View commit details
    Browse the repository at this point in the history

  12. load recordings from manifest if it is path

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    3eaf62f View commit details
    Browse the repository at this point in the history

  13. create a detector register in the base class

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    fe0755f View commit details
    Browse the repository at this point in the history

  14. remove Protocol for compatibility with 3.7

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    480ac67 View commit details
    Browse the repository at this point in the history

  15. rename activity_detection to detect_activity

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    6a6971c View commit details
    Browse the repository at this point in the history

  16. fix RecordingSet loading from file

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    863a65f View commit details
    Browse the repository at this point in the history

  17. update docs

    @rilshok
    rilshok committed Sep 29, 2023
    Configuration menu
    Copy the full SHA
    fe0599b View commit details
    Browse the repository at this point in the history

Commits on Sep 30, 2023

  1. refactor the trimming procedure

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    1e8c718 View commit details
    Browse the repository at this point in the history

  2. add docstrings and modify exceptions

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    cdd0a3a View commit details
    Browse the repository at this point in the history

  3. add base test for tim inactivity

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    d415096 View commit details
    Browse the repository at this point in the history

  4. add test for trim-inactivity workflow

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    44a960c View commit details
    Browse the repository at this point in the history

  5. process only the part of audio defined in the Cut

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    911e63c View commit details
    Browse the repository at this point in the history

  6. expand testing

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    f998797 View commit details
    Browse the repository at this point in the history

  7. the file with the record has the same name as cut

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    874d3c7 View commit details
    Browse the repository at this point in the history

  8. fix external interval protection error

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    2e1e1de View commit details
    Browse the repository at this point in the history

  9. resampling before feeding to the detector

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    d21b599 View commit details
    Browse the repository at this point in the history

  10. always use spawn context; don't use dict for model

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    c0151af View commit details
    Browse the repository at this point in the history

  11. test known detectors

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    6e5b561 View commit details
    Browse the repository at this point in the history

  12. do not cover path checking

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    17adbab View commit details
    Browse the repository at this point in the history

  13. save the file to disc in the standard way

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    162b4f8 View commit details
    Browse the repository at this point in the history

  14. use class._models:dict to store models

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    ea75a05 View commit details
    Browse the repository at this point in the history

  15. create output dir if not exists

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    8a16558 View commit details
    Browse the repository at this point in the history

  16. update docs

    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    2992767 View commit details
    Browse the repository at this point in the history

  17. Merge branch 'master' of github.com:lhotse-speech/lhotse into feature… 

    …/distill-cutset
    @rilshok
    rilshok committed Sep 30, 2023
    Configuration menu
    Copy the full SHA
    4fcadc7 View commit details
    Browse the repository at this point in the history

Commits on Oct 1, 2023

  1. use the detector_name variable instead of name

    @rilshok
    rilshok committed Oct 1, 2023
    Configuration menu
    Copy the full SHA
    51d53b2 View commit details
    Browse the repository at this point in the history

  2. use csv instead of pandas

    @rilshok
    rilshok committed Oct 1, 2023
    Configuration menu
    Copy the full SHA
    0c7b75c View commit details
    Browse the repository at this point in the history