Elly Knight
@ellycknight.bsky.social
Birds, acoustics, nightjars, movement
Quantitative Ecologist at Alberta Biodiversity Monitoring Institute & Boreal Avian Modelling Project
Parent of 3 (+ doggo), fibre lover, cyclist
She/her
Quantitative Ecologist at Alberta Biodiversity Monitoring Institute & Boreal Avian Modelling Project
Parent of 3 (+ doggo), fibre lover, cyclist
She/her
Awesome David! Let us know if you need any help getting set up!
May 15, 2025 at 2:42 PM
Awesome David! Let us know if you need any help getting set up!
Want to try it?
Find open-access code and models for HawkEars on Github at github.com/jhuus/HawkEars
Watch for future integration into WildTrax at wildtrax.ca
#bioacoustics 4/4
Find open-access code and models for HawkEars on Github at github.com/jhuus/HawkEars
Watch for future integration into WildTrax at wildtrax.ca
#bioacoustics 4/4
GitHub - jhuus/HawkEars: Scan audio recordings for bird sounds and generate Audacity label files
Scan audio recordings for bird sounds and generate Audacity label files - GitHub - jhuus/HawkEars: Scan audio recordings for bird sounds and generate Audacity label files
github.com
May 14, 2025 at 10:16 PM
Want to try it?
Find open-access code and models for HawkEars on Github at github.com/jhuus/HawkEars
Watch for future integration into WildTrax at wildtrax.ca
#bioacoustics 4/4
Find open-access code and models for HawkEars on Github at github.com/jhuus/HawkEars
Watch for future integration into WildTrax at wildtrax.ca
#bioacoustics 4/4
HawkEars will be of interest to practitioners studying acoustic wildlife in Canada and the northern United States, but the open-source code can also be used for developing classifiers for other regions.
#bioacoustics 3/n
#bioacoustics 3/n
May 14, 2025 at 10:15 PM
HawkEars will be of interest to practitioners studying acoustic wildlife in Canada and the northern United States, but the open-source code can also be used for developing classifiers for other regions.
#bioacoustics 3/n
#bioacoustics 3/n
We used an embedding-based search to generate strong labels from weakly-labelled open-access training datasets. Our regional model outperformed existing global models, BirdNET, and Perch, by substantial margins, setting benchmarks for state-of-the-art classification performance.
#bioacoustics 2/n
#bioacoustics 2/n
May 14, 2025 at 10:14 PM
We used an embedding-based search to generate strong labels from weakly-labelled open-access training datasets. Our regional model outperformed existing global models, BirdNET, and Perch, by substantial margins, setting benchmarks for state-of-the-art classification performance.
#bioacoustics 2/n
#bioacoustics 2/n
I totally missed this way back when, thank you Ryan!
April 11, 2025 at 10:07 PM
I totally missed this way back when, thank you Ryan!