The repository has been divided into 2 separate sections.
-
The first task i.e. the audio captioning task is present in task6a/.
a. Experiments with the baseline repo can be found inside task6a/baseline_exps. The non-modified repository of the challenge can be found here: https://github.com/felixgontier/dcase-2023-baseline
b. Experiments with the best submission from last year's challenge can be found inside task6a/exps. Please see the additional_run_notes in order to get the code running.
c. The evaluation script for this task can be found inside task6a/fense
-
The second task i.e. the audio retrieval task is present in task6b/. The base repository can be found here: https://github.com/xieh97/dcase2023-audio-retrieval
a. Run the corresponding main files and change the conf.yaml file to consider the required embeddings.
b. Additional instructions to run the repository can be found in task6b/README.md.