François Chollet on Twitter: "Many runtimes other than Python TensorFlow understand the Keras graph-of-layers format, such as TF.js, CoreML, DeepLearning4J... A high level, human-readable saving format is much easier to implement for third-party platforms."

Country	Code	For customers of
United States	40404	(any)
Canada	21212	(any)
United Kingdom	86444	Vodafone, Orange, 3, O2
Brazil	40404	Nextel, TIM
Haiti	40404	Digicel, Voila
Ireland	51210	Vodafone, O2
India	53000	Bharti Airtel, Videocon, Reliance
Indonesia	89887	AXIS, 3, Telkomsel, Indosat, XL Axiata
Italy	4880804	Wind
3424486444	Vodafone
» See SMS short codes for other countries

François Chollet‏Verified account @fchollet Jul 12

This builds the following graph:pic.twitter.com/WQzvf6bJZh

1 reply 1 retweet 24 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

Now, of course, you could also define such a model as a Python class. It would then look like this:pic.twitter.com/2oU0vjOIHe

3 replies 1 retweet 26 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

But there are several key advantages of the Functional approach over the subclassing approach: 1. Your model has known inputs shapes. 2. You get access to the internal connectivity graph. 3. The model is a data structure, not a piece of bytecode. Let's see what these are about.

2 replies 0 retweets 33 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

1. Because the model has known input shapes, it's capable of running input validation checks, for easy debugging:pic.twitter.com/1B8E7GXmK1

1 reply 1 retweet 19 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

Further, it's even capable of standardizing inputs to what it expects: if you pass data of shape (batch_size,) to a model that expects (batch_size, 1), it will just reshape it. Likewise for dtype conversion (e.g. float64 will get converted to float32).

1 reply 0 retweets 15 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

2. You get access to the internal connectivity graph. This means you can plot the model, for instance. This is great for debugging. Like this:pic.twitter.com/ZnG6ym9yei

1 reply 0 retweets 38 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

Having access to internal nodes also means you can access an intermediate layer output and leverage it in a new model. This is a killer feature for feature extraction, fine-tuning, and ensembling. Let's add an extra output to the model above:pic.twitter.com/gCxafm21UF

1 reply 0 retweets 27 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

3. The model is a data structure, not a piece of bytecode. This means it can be cleanly serialized and deserialized -- even across platforms. keras.Model.from_config(functional_model.get_config()) reconstructs the exact same model as the original.

1 reply 1 retweet 24 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

If your model is a Python subclass, to serialize it you could either: a. Pickle the bytecode -- which it completely unsafe, won't work for production, and won't work across platforms

2 replies 2 retweets 15 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

b. Save it as a SavedModel -- which is a form of one-way export (of the TF graph) and won't let you reconstruct the exact same Python object. A graph of layers is a data structure; defining and saving it as a data structure is the intuitive thing to do.

2 replies 3 retweets 20 likes

Show this thread

François Chollet‏Verified account @fchollet Jul 12

Many runtimes other than Python TensorFlow understand the Keras graph-of-layers format, such as TF.js, CoreML, DeepLearning4J... A high level, human-readable saving format is much easier to implement for third-party platforms.

12:39 PM - 12 Jul 2021