Merge pull request #76 from azavea/lf/0.11

Release v0.11

Author: lewfish

README.md

@@ -4,6 +4,7 @@ This repository contains examples of using Raster Vision on open datasets.
 
 Note: The `master` branch of this examples repo should be used in conjunction with the `master` branch (or `latest` Docker image tag) of [Raster Vision](https://github.com/azavea/raster-vision)
 which contains the latest changes. For versions of this examples repo that correspond to stable, released versions of Raster Vision, see:
+* [0.11](https://github.com/azavea/raster-vision-examples/tree/0.11)
 * [0.10](https://github.com/azavea/raster-vision-examples/tree/0.10)
 * [0.9](https://github.com/azavea/raster-vision-examples/tree/0.9)
 
@@ -110,7 +111,7 @@ The dataset is stored on AWS S3 at `s3://spacenet-dataset`. You will need an AWS
 
 Optional: to run this example with the data stored locally, first copy the data using something like the following inside the container.
 ```
-aws s3 sync s3://spacenet-dataset/AOI_1_Rio/ /opt/data/spacenet-dataset/AOI_1_Rio/
+aws s3 sync s3://spacenet-dataset/AOIs/AOI_1_Rio/ /opt/data/spacenet-dataset/AOIs/AOI_1_Rio/
 ```
 
 #### Step 2: Run the Jupyter Notebook

docker/build

@@ -26,15 +26,15 @@ then
     then
         if [ -z "$RASTER_VISION_REPO" ]
         then
-            BASE_IMAGE="quay.io/azavea/raster-vision:tf-cpu-latest"
+            BASE_IMAGE="quay.io/azavea/raster-vision:tf-cpu-0.11"
         else
             BASE_IMAGE="raster-vision-tf-cpu"
         fi
         docker build --build-arg BASE_IMAGE="$BASE_IMAGE" -t raster-vision-examples-tf-cpu -f Dockerfile .
 
         if [ -z "$RASTER_VISION_REPO" ]
         then
-            BASE_IMAGE="quay.io/azavea/raster-vision:tf-gpu-latest"
+            BASE_IMAGE="quay.io/azavea/raster-vision:tf-gpu-0.11"
         else
             BASE_IMAGE="raster-vision-tf-gpu"
         fi
@@ -45,7 +45,7 @@ then
     then
         if [ -z "$RASTER_VISION_REPO" ]
         then
-            BASE_IMAGE="quay.io/azavea/raster-vision:pytorch-latest"
+            BASE_IMAGE="quay.io/azavea/raster-vision:pytorch-0.11"
         else
             BASE_IMAGE="raster-vision-pytorch"
         fi

examples/spacenet/vegas/all.py

@@ -74,11 +74,11 @@ def get_class_id_to_filter(self):
 
 class VegasBuildings(SpacenetConfig):
     def __init__(self, raw_uri):
-        self.base_dir = 'SpaceNet_Buildings_Dataset_Round2/spacenetV2_Train/AOI_2_Vegas'
-        self.raster_dir = 'RGB-PanSharpen/'
-        self.label_dir = 'geojson/buildings/'
-        self.raster_fn_prefix = 'RGB-PanSharpen_AOI_2_Vegas_img'
-        self.label_fn_prefix = 'buildings_AOI_2_Vegas_img'
+        self.base_dir = 'spacenet/SN2_buildings/train/AOI_2_Vegas'
+        self.raster_dir = 'PS-RGB'
+        self.label_dir = 'geojson_buildings'
+        self.raster_fn_prefix = 'SN2_buildings_train_AOI_2_Vegas_PS-RGB_img'
+        self.label_fn_prefix = 'SN2_buildings_train_AOI_2_Vegas_geojson_buildings_img'
         super().__init__(raw_uri)
 
     def get_class_map(self):

examples/spacenet/vegas/simple_segmentation.py

@@ -30,15 +30,15 @@ def exp_main(self, raw_uri, root_uri, test=False):
         """
         # Specify the location of the raw data
         base_uri = join(
-            raw_uri, 'SpaceNet_Buildings_Dataset_Round2/spacenetV2_Train/AOI_2_Vegas')
+            raw_uri, 'spacenet/SN2_buildings/train/AOI_2_Vegas')
         # The images and labels are in two separate directories within the base_uri
-        raster_uri = join(base_uri, 'RGB-PanSharpen')
-        label_uri = join(base_uri, 'geojson/buildings')
+        raster_uri = join(base_uri, 'PS-RGB')
+        label_uri = join(base_uri, 'geojson_buildings')
         # The tiff (raster) and geojson (label) files have have a naming convention of
         # '[prefix]_[image id].geojson.' The prefix indicates the type of data and the
         # image id indicates which scene each is associated with.
-        raster_fn_prefix = 'RGB-PanSharpen_AOI_2_Vegas_img'
-        label_fn_prefix = 'buildings_AOI_2_Vegas_img'
+        raster_fn_prefix = 'SN2_buildings_train_AOI_2_Vegas_PS-RGB_img'
+        label_fn_prefix = 'SN2_buildings_train_AOI_2_Vegas_geojson_buildings_img'
         # Find all of the image ids that have associated images and labels. Collect
         # these values to use as our scene ids.
         label_paths = list_paths(label_uri, ext='.geojson')

examples/test.py

@@ -37,6 +37,7 @@
             'raw_uri': 's3://raster-vision-raw-data/isprs-potsdam',
             'processed_uri': 's3://raster-vision-lf-dev/examples/cowc-potsdam/processed-data',
         },
+        'rv_profile': 'pytorch',
     },
     {
         'key': 'potsdam-semantic-segmentation-pytorch',
@@ -48,7 +49,8 @@
         'remote': {
             'raw_uri': 's3://raster-vision-raw-data/isprs-potsdam',
             'processed_uri': 's3://raster-vision-lf-dev/examples/potsdam/processed-data',
-        }
+        },
+        'rv_profile': 'pytorch',
     },
     {
         'key': 'potsdam-semantic-segmentation-tf',
@@ -75,6 +77,7 @@
             'raw_uri': 's3://spacenet-dataset/',
             'processed_uri': 's3://raster-vision-lf-dev/examples/spacenet/rio/processed-data',
         },
+        'rv_profile': 'pytorch',
     },
     {
         'key': 'spacenet-rio-chip-classification-tf',
@@ -102,7 +105,8 @@
             'processed_uri': 's3://raster-vision-lf-dev/examples/spacenet/vegas/processed-data',
         },
         'extra_args': [['target', 'buildings'],
-                       ['task_type', 'semantic_segmentation']]
+                       ['task_type', 'semantic_segmentation']],
+        'rv_profile': 'pytorch',
     },
     {
         'key': 'spacenet-vegas-roads-semantic-segmentation-pytorch',
@@ -116,7 +120,8 @@
             'processed_uri': 's3://raster-vision-lf-dev/examples/spacenet/vegas/processed-data',
         },
         'extra_args': [['target', 'roads'],
-                       ['task_type', 'semantic_segmentation']]
+                       ['task_type', 'semantic_segmentation']],
+        'rv_profile': 'pytorch',
     },
     {
         'key': 'xview-object-detection-tf',
@@ -140,7 +145,7 @@ def run_experiment(exp_cfg, root_uri, test=True, remote=False, commands=None):
     rv_profile = exp_cfg.get('rv_profile')
     if rv_profile is not None:
         cmd += ['-p', rv_profile]
-    cmd += ['run', 'aws_batch' if remote else 'local', '-e', exp_cfg['module']]
+    cmd += ['run', 'aws_batch' if remote else 'inprocess', '-e', exp_cfg['module']]
     cmd += ['-a', 'raw_uri', uris['raw_uri']]
     if 'processed_uri' in uris:
         cmd += ['-a', 'processed_uri', uris['processed_uri']]
@@ -250,4 +255,4 @@ def collect_eval_dir(root_uri):
 
 
 if __name__ == '__main__':
-    test()
+    test()