assets: draft for tfjs model for standard_v2_0

[reyammer]

This is an attempt to convert standard_v2_0 to tfjs format.

The model is untested.

Steps:

• Convert .keras model to saved_model format: tf.keras.models.save_model(model , './savedmodel/', save_format='tf')
• Create a fresh python virtual environment with python<=3.11 (tensorflowjs breaks with python 3.12)
• pip install tensorflowjs
• tensorflowjs_converter --input_format=tf_saved_model --output_format=tfjs_graph_model ./savedmodel ./magika_tfjs

Notes:

• the conversion to saved model is needed because we use custom layers, .keras does not store them, and I could not find a way to add custom layers to the scope of the tensorflowjs_converter CLI script.