Your challenge in this competition is to identify which birds are calling in long recordings made in Kenya. This is an important task for scientists who monitor bird populations for conservation purposes. More accurate solutions could enable more comprehensive monitoring. This year, your notebook must also complete inference in [a more constrained time frame](https://www.kaggle.com/competitions/birdclef-2023/overview/code-requirements). This will make it easier to deploy winning models for on the ground conservation efforts where efficiency is at a premium.

This competition uses a hidden test. When your submitted notebook is scored, the actual test data (including a sample submission) will be made available to your notebook.


##Files

**train_audio/** The training data consists of short recordings of individual bird calls generously uploaded by users of [xenocanto.org](https://www.xeno-canto.org/). These files have been downsampled to 32 kHz where applicable to match the test set audio and converted to the ogg format. The training data should have nearly all relevant files; we expect there is no benefit to looking for more on [xenocanto.org](https://www.xeno-canto.org/).


**test_soundscapes/** When you submit a notebook, the **test_soundscapes** directory will be populated with approximately 200 recordings to be used for scoring. They are 10 minutes long and in ogg audio format. The file names are randomized. It should take your submission notebook approximately five minutes to load all of the test soundscapes.


**train_metadata.csv**  A wide range of metadata is provided for the training data. The most directly relevant fields are:

*   `primary_label` - a code for the bird species. You can review detailed information about the bird codes by appending the code to `https://ebird.org/species/`, such as `https://ebird.org/species/amecro` for the American Crow.
* `latitude` & `longitude`: coordinates for where the recording was taken. Some bird species may have local call 'dialects,' so you may want to seek geographic diversity in your training data.
*   `author` - The user who provided the recording.
* `filename`: the name of the associated audio file.


**sample_submission.csv** A valid sample submission. 
* `row_id`: A slug of `[soundscape_id]_[end_time]` for the prediction.
* `[bird_id]`: There are 264 bird ID columns. You will need to predict the probability of the presence of each bird for each row.

**eBird_Taxonomy_v2021.csv** - Data on the relationships between different species.