We do not need the metadata file because an audio sample’s label is inscribed in its filename. Parsing audio data into TFRecordsĪs a word beforehand, if this is your first time working with TFRecords, TensorFlow’s native format for efficient data storage, or you need a refresher, have a look at this Google Colab notebook and an accompanying description here. The first part is used to identify the ID of the source file, LABEL is for the class label, and TAKE and SLICE are used to distinguish between multiple samples taken from the same original file. All files are named according to the following scheme: Within this folder, the files are pre-arranged into ten folders, fold1 to fold10. You now have two folders, audio and metadata. As mentioned, you can use the data free of charge, but only for non-commercial projects - for this explanatory blog post showcasing data parsing and related concepts, things are fine.ĭepending on your connection, the download might take a few minutes. Go to the dataset’s web page and click on the download link at the bottom to download the dataset. Thus, if you’re looking for a smaller, more-accessible audio dataset, then have a look at ESC50. Each file might have a different sampling rate, bit-depth, and number of channels. The files come pre-arranged in ten folds and are stored in WAVE format. At most 4 seconds long, each sample belongs to one of 10 classes. The UrbanSound8K dataset (available free of charge for non-commercial use only CC BY-NC 3.0) contains 8732 audio files of varying duration. If you are completely new to this format, you can get a hands-on recap here, which also covers how we can parse text and image data. Here, we’ll go over parsing the UrbanSound8K dataset, store it in the TFRecord format, and, finally, iterate over its samples. Thankfully, getting other datasets into the TFRecord format is not that hard once you’ve done it a few times. Though many datasets are readily available, such as MNIST or CIFAR, you sometimes have to implement custom ones yourself.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |