Content The dataset consists of two files, training and validation. Each folder contains 10 subforders labeled as n0~n9, each corresponding a species form [Wikipedia's monkey cladogram][1]. Images are 400x300 px or larger and JPEG format (almost 1400 images). Images were downloaded with help of the [googliser][2] open source code. Label mapping: n0, alouatta_palliata n1, erythrocebus_patas n2, cacajao_calvus n3, macaca_fuscata n4, cebuella_pygmea n5, cebus_capucinus n6, mico_argentatus n7, saimiri_sciureus n8, aotus_nigriceps n9, trachypithecus_johnii - For more information on the monkey species and number of images per class make sure to check `monkey_labels.txt` file. Aim This dataset is intended as a test case for fine-grain classification tasks, perhaps best used in combination with transfer learning. Hopefully someone can help us expand the number of classes or number of images. Acknowledgements Thanks to Romain Renard for his help with the code implementation. Also, thanks to Gustavo Montoya, Jacky Zhang and Sofia Loaiciga for their help with the dataset curation. Notes Some demo code for usage of the dataset in combination with Keras can be found in this [repo][3]. [1]: https://en.wikipedia.org/wiki/Monkey [2]: https://github.com/teracow/googliser [3]: https://github.com/slothkong/CNN_classification_10_monkey_species