summarize¶
Generate a text summary of a model.
Additional Documentation¶
Usage¶
Usage: mltk summarize [OPTIONS] <model>
Generate a summary of a model
If a .h5 file is provided or a MLTK model name/archive/script and *no*
--tflite option,
then print a summary of the Keras Model.
If a .tflite file is provided or a MLTK model name/archive/script and the
--tflite option,
then print a summary of the .tflite model.
Use the --build option if the model has not been previously trained.
For more details see:
https://siliconlabs.github.io/mltk/docs/guides/model_summary
----------
Examples
----------
# Print a summary of pre-trained Keras model
mltk summarize audio_example1
# Print a summary of pre-trained TF-Lite model
mltk summarize audio_example1 --tflite
# Generate a .tflite then print a summary
# In this case, the model need not be previously trained
mltk summarize audio_example1 --build --tflite
# Print of summary of the given .tflite
mltk summarize some/path/my_model.tflite
# Print of summary of the given model archive's .tflite
mltk summarize some/path/my_model.mltk.zip --tflite
Arguments
* model <model> One of the following:
- Path to .tflite model file
- Path to .h5 model file
- Name of MLTK model
- Path to trained model's archive (.mltk.zip)
- Path to MLTK model's python script
[default: None]
[required]
Options
--tflite Summarize the .tflite model file in the MLTK
model's archive, or if the --build option is
given, generate .tflite file before summarizing
--build -b Build the model rather than loading from a
pre-trained file in the MLTK model archive
--output -o <path> File path of generated summary file. If omitted,
the summary is printed to console
[default: None]
--verbose -v Enable verbose console logs
--help Show this message and exit.