The SDK provides a Jupyter Notebook to deserialize for inference result.
By running the deserialize.ipynb, you can deserialize the inference result.
Prepare a schema file for the serialized data and a file containing the serialized data in any directory under the deserialize
directory.
In the VS Code TERMINAL tab, run following.
$ ./tutorials/_common/deserialize/build.sh
After a few minutes, a docker image is created for deserialization.
Place setting file (./configuration.json
file) for deserialize.
- configuration.json
{ "schema_file" : "", "serialized_file" : "", "input_type" : "json", "output_dir" : "./output/" }
NOTE
See 4. Edit settings for details on the parameter.
Edit the parameters in configuration.json.
The parameters required to run this notebook are :
Setting | Description | Range | Required/Optional | Remarks |
---|---|---|---|---|
schema_file |
Path to schema file for serialized data | Absolute path or relative path from configuration.json/Notebook (*.ipynb) | Required | |
serialized_file |
Path to file containing the serialized data | Absolute path or relative path from configuration.json/Notebook (*.ipynb) | Required | |
input_type |
Type of serialized_file |
"binary" or "json" | Required | |
output_dir |
Path to output directory for deserialized file (a new directory will be created if it does not exist) |
Absolute path or relative path from configuration.json/Notebook (*.ipynb) | Optional | If omitted or given an empty string, set path same as configuration.json/Notebook (*.ipynb). The format of the output filename is " serialized_file without extension" + ".json". |
Open notebook and run the cells.
If successful, deserialized file will be output in output_dir
.
You can run all cells at once, or you can run the cells one by one.
- FlatBuffers
The version of FlatBuffers used in "Vision and Sensing Application SDK" is 23.1.21.