update vessel detection docs and Dockerfiles to point to new checkpoints

Author: favyen2

docs/landsat_vessels.md

@@ -27,8 +27,8 @@ First, download the detector, classifier, and attribute model checkpoints to the
     mkdir -p project_data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/
     wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/best.ckpt -O project_data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/best.ckpt
 
-    mkdir -p project_data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/
-    wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/best.ckpt -O project_data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/best.ckpt
+    mkdir -p project_data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/
+    wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/best.ckpt -O project_data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/best.ckpt
 
 The easiest way to apply the model is using the prediction pipeline in `rslp/landsat_vessels/predict_pipeline.py`. You can download the Landsat scene files, e.g. from USGS EarthExplorer or AWS, and then create a configuration file for the prediction pipeline, here is an example:
 

docs/sentinel1_vessels.md

@@ -16,8 +16,8 @@ First, download the model checkpoints to the `RSLP_PREFIX` directory.
     mkdir -p project_data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/
     wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/best.ckpt -O project_data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/best.ckpt
 
-    mkdir -p project_data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/
-    wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/best.ckpt -O project_data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/best.ckpt
+    mkdir -p project_data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/
+    wget https://storage.googleapis.com/ai2-rslearn-projects-data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/best.ckpt -O project_data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/best.ckpt
 
 The easiest way to apply the model is using the prediction pipeline in
 `rslp/sentinel1_vessels/predict_pipeline.py`. It accepts a Sentinel-1 scene ID and

rslp/landsat_vessels/Dockerfile

@@ -20,9 +20,9 @@ RUN mkdir -p /project_data/projects/landsat_vessels/data_20240924_model_20240924
     mkdir -p /project_data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/ && \
     wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/best.ckpt \
         -O /project_data/projects/rslearn-landsat-recheck/phase123_20240919_01_copy/best.ckpt && \
-    mkdir -p /project_data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/ && \
-    wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/best.ckpt \
-        -O /project_data/projects/landsat_vessel_attribute/data_20260330_olmoearth_base_freeze_01/best.ckpt
+    mkdir -p /project_data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/ && \
+    wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/best.ckpt \
+        -O /project_data/projects/2026_04_22_landsat_vessel_attribute/olmoearth_base_freeze_newdata_01/best.ckpt
 ENV RSLP_PREFIX=/project_data
 
 # Run app.py when the container launches

rslp/sentinel1_vessels/Dockerfile

@@ -17,9 +17,9 @@ ENV MARINE_INFRA_PATH=/opt/rslearn_projects/marine_infrastructure.geojson
 RUN mkdir -p /project_data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/ && \
     wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/best.ckpt \
         -O /project_data/projects/sentinel1_vessels/data_20250521_model_20250530_satlaspretrain_unfreeze4_13/best.ckpt && \
-    mkdir -p /project_data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/ && \
-    wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/best.ckpt \
-        -O /project_data/projects/sentinel1_vessel_attribute/data_20260330_swinb_01/best.ckpt
+    mkdir -p /project_data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/ && \
+    wget -q https://storage.googleapis.com/ai2-rslearn-projects-data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/best.ckpt \
+        -O /project_data/projects/2026_04_22_sentinel1_vessel_attribute/swinb_newdata_01/best.ckpt
 ENV RSLP_PREFIX=/project_data
 
 # Run app.py when the container launches