everything AI about roofs. 🏛️
🔷 datasets 🔷 notebooks 🔷 semseg 🔷 inference 🔷 sagemaker 🔷
pip install roofAI > roofAI help
cloudwatch browse \
[endpoint] \
[<endpoint-name>]
. browse endpoint on cloudwatch.
* default endpoint: endpoint-model-2023-12-03-11-24-39-75649-pytorch
roofAI conda activate [sagemaker|semseg]
. activate conda environment.
roofAI conda create [~validate,~recreate,sagemaker|semseg]
. create conda environment.
roofAI conda environment_name [sagemaker|semseg]
. return conda environment_name
roofAI conda validate
. validate conda environment.
roofAI inference create \
[dryrun,model] \
[.|<object-name>] \
[--verbose 1] \
[--verify 0]
. create inference model.
roofAI inference create \
[dryrun,endpoint_config,suffix=<suffix>] \
[.|<object-name>] \
[--verbose 1] \
[--verify 0]
. create inference endpoint config.
roofAI inference create \
[dryrun,endpoint,config_suffix=<suffix>,suffix=<suffix>] \
[.|<object-name>] \
[--verbose 1] \
[--verify 0]
. create inference endpoint.
roofAI inference delete \
[dryrun,model|endpoint_config|endpoint] \
<name> \
[--verbose 1]
. delete inference object.
roofAI inference describe \
[dryrun,endpoint] \
<name> \
[--verbose 1]
. describe inference endpoint.
roofAI inference invoke \
[~download,dryrun,profile=FULL|DECENT|QUICK|DEBUG|VALIDATION,upload] \
[-|<endpoint-name>] \
[..|<dataset-object-name>] \
[-|<prediction-object-name>] \
[--verbose 1]
. <dataset-object-name> -> inference endpoint -> <prediction-object-name>.
* default endpoint: endpoint-model-2023-12-03-11-24-39-75649-pytorch
roofAI inference list \
[dryrun,model|endpoint_config|endpoint,contains=<string>] \
[--verbose 1]
. list inference objects.
roofAI inference pull \
[dryrun]
. pull the inference image.
QGIS seed [screen]
. seed 🌱 QGIS.
QGIS expressions pull
. pull QGIS expressions.
QGIS expressions push [push]
. push QGIS expressions.
📂 /Users/kamangir/Library/Application Support/QGIS/QGIS3/profiles/default/python/expressions
📂 /Users/kamangir/git/roofAI/roofAI/QGIS/expressions
QGIS serve[r] [start]
. start QGIS server.
semseg list
. list registered semseg models.
semseg predict \
[device=cpu|cuda,~download,dryrun,profile=FULL|DECENT|QUICK|DEBUG|VALIDATION,upload] \
[...|<model-object-name>] \
[..|<dataset-object-name>] \
[-|<prediction-object-name>]
. semseg[<model-object-name>].predict(<dataset-object-name>) -> <prediction-object-name>.
semseg train \
[device=cpu|cuda,~download,dryrun,profile=FULL|DECENT|QUICK|DEBUG|VALIDATION,register,suffix=<v1>,upload] \
<dataset-object-name> \
<model-object-name> \
[--activation <sigmoid>] \
[--classes <one+two+three+four>] \
[--encoder_name <se_resnext50_32x4d>] \
[--encoder_weights <imagenet>]
. semseg.train(<dataset-object-name>) -> <model-object-name>.
roofAI dataset ingest \
[source=AIRS,dryrun,open,register,suffix=<v1>,upload,target=sagemaker|torch] \
<object-name> \
[--test_count <10>] \
[--train_count <10>] \
[--val_count <10>]
. ingest AIRS -> <object-name>.
roofAI dataset ingest \
[source=CamVid,dryrun,open,register,suffix=<v1>,upload] \
<object-name>
. ingest CamVid -> <object-name>.
roofAI dataset review \
[download,dryrun,open] \
<object-name> \
[--count <1>] \
[--index <index>] \
[--subset <subset>]
. review <object-name>.
To use on AWS SageMaker replace <plugin-name> with roofAI and follow these instructions.