{ "id": 175419, "title": "Offline models for SED", "url": "/competitions/birdsong-recognition/discussion/175419", "author_name": "Mighty Rains", "post_date": "2020-08-18T06:14:42.396000", "votes": 4, "comment_count": 0, "views": 0, "content": "

@hidehisaarai1213 made a great notebook introducing SED. Since the competition needs to be run offline and the PANNs weights are not downloadable from an API inside, I made a notebook which has the all the required libraries (torchlibrosa, etc) and the pre-trained weights too, so it becomes a bit easy for offline use.

\n

Here it is: https://www.kaggle.com/mightyrains/panns-datasets

", "messages": [ { "id": 975074, "postDate": "2020-08-18T06:14:42.397Z", "content": "

@hidehisaarai1213 made a great notebook introducing SED. Since the competition needs to be run offline and the PANNs weights are not downloadable from an API inside, I made a notebook which has the all the required libraries (torchlibrosa, etc) and the pre-trained weights too, so it becomes a bit easy for offline use.

\n

Here it is: https://www.kaggle.com/mightyrains/panns-datasets

", "rawMarkdown": "@hidehisaarai1213 made a great [notebook](https://www.kaggle.com/hidehisaarai1213/introduction-to-sound-event-detection) introducing SED. Since the competition needs to be run offline and the PANNs weights are not downloadable from an API inside, I made a notebook which has the all the required libraries (torchlibrosa, etc) and the pre-trained weights too, so it becomes a bit easy for offline use.\n\nHere it is: [https://www.kaggle.com/mightyrains/panns-datasets](https://www.kaggle.com/mightyrains/panns-datasets)", "votes": 4 } ], "comments": [], "raw_markdown_by_id": { "975074": "@hidehisaarai1213 made a great [notebook](https://www.kaggle.com/hidehisaarai1213/introduction-to-sound-event-detection) introducing SED. Since the competition needs to be run offline and the PANNs weights are not downloadable from an API inside, I made a notebook which has the all the required libraries (torchlibrosa, etc) and the pre-trained weights too, so it becomes a bit easy for offline use.\n\nHere it is: [https://www.kaggle.com/mightyrains/panns-datasets](https://www.kaggle.com/mightyrains/panns-datasets)" } }