strongly label bamboo eating from the collar data

One issue with the bamboo eating class in the panda dataset is the segments were generated randomly from minute long clips of panda eating. because the main goal was chirps, this seemed ok for a noise class, but for a bamboo eating classifier, this data needs to be strongly labeled. will bring it into labelstudio and label them there.