From 0f92e1a43899ab5f5c21ce9d27ef42492d9c12f2 Mon Sep 17 00:00:00 2001 From: bhandarivijay Date: Tue, 6 Jan 2026 07:00:09 +0000 Subject: [PATCH] chore: Migrate gsutil usage to gcloud storage --- tensorflow-planespotting/README.md | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/tensorflow-planespotting/README.md b/tensorflow-planespotting/README.md index 06863e2..ee3f884 100644 --- a/tensorflow-planespotting/README.md +++ b/tensorflow-planespotting/README.md @@ -76,11 +76,11 @@ This also allows better shuffling, which is important for the final accuracy. Both formats are available in a public Cloud Storage bucket. You can list them using `gsutil`: ```bash # large aerial photographs and ROI files: - gsutil ls -l gs://planespotting-data-public/USGS_public_domain_photos - gsutil ls -l gs://planespotting-data-public/USGS_public_domain_photos_eval + gcloud storage ls --long gs://planespotting-data-public/USGS_public_domain_photos + gcloud storage ls --long gs://planespotting-data-public/USGS_public_domain_photos_eval # 256x256 pixel tiles and adjusted ROI files: - gsutil ls -l gs://planespotting-data-public/tiles_from_USGS_photos - gsutil ls -l gs://planespotting-data-public/tiles_from_USGS_photos_eval + gcloud storage ls --long gs://planespotting-data-public/tiles_from_USGS_photos + gcloud storage ls --long gs://planespotting-data-public/tiles_from_USGS_photos_eval ``` You can run the data generation alone using this script. Make sure you provide your @@ -158,8 +158,8 @@ run it from a local web browser: cd webui-mark # copy the existing data, add yours if you have any mkdir USGS_public_domain_airports -gsutil -m cp gs://planespotting-data-public/USGS_public_domain_photos/* USGS_public_domain_airports -gsutil -m cp gs://planespotting-data-public/USGS_public_domain_photos_eval/* USGS_public_domain_airports +gcloud storage cp gs://planespotting-data-public/USGS_public_domain_photos/* USGS_public_domain_airports +gcloud storage cp gs://planespotting-data-public/USGS_public_domain_photos_eval/* USGS_public_domain_airports # run from local server (any port will do) python -m SimpleHTTPServer 8000 ``` @@ -184,4 +184,4 @@ in the `serving_input_fn` function so that it happens on the fly when the model is deployed on AI Platform. It is a good example of how deployed models can run preprocessing on their input data directly on AI Platform while serving. Code here: [trainer/train.py](trainer/train.py) function `serving_input_fn`. - + \ No newline at end of file