I am trying to prepare the COCO dataset. on link https://cloud.google.com/tpu/docs/tutorials/retinanet#prepare_the_coco_dataset it is given "The tpu/tools/datasets/download_and_preprocess_coco.sh script converts the COCO dataset into a set of TFRecords that the training application expects"
When i try it i am getting -bash: tpu/tools/datasets/download_and_preprocess_coco.sh: No such file or directory
How can i solve it
Assuming that you are using one of the ml-images provided by Google Cloud, that script should reside in the /usr/share/tpu/ directory, which is a copy of the github repository at https://github.com/tensorflow/tpu.
I ran the following command and got the following result on a brand-new instance created from the 1.12 image.
$ file /usr/share/tpu/tools/datasets/download_and_preprocess_coco.sh
/usr/share/tpu/tools/datasets/download_and_preprocess_coco.sh: Bourne-Again shell script, ASCII text executable
Please let me know if this works, or if you run into any further difficulties.
