Adapting README content

README.md

@@ -13,7 +13,7 @@ metrics:
 ---
 
 ## Model name: mobilenet_v3_small_100_224
-## Description taken from [TFHub](https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/classification/5)
+## Description adapted from [TFHub](https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/classification/5)
 
 # Overview
 
@@ -32,7 +32,7 @@ For a quick comparison between these variants, please refer to the following tab
 |Small|1.0|67.5|15.8|19.4|14.4|
 |Small|0.75|65.4|12.8|15.9|11.6|
 
-This TF Hub model uses the TF-Slim implementation of [`mobilenet_v3`](https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v3.py) as a small network with a depth multiplier of 1.0.
+This model uses the TF-Slim implementation of [`mobilenet_v3`](https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v3.py) as a small network with a depth multiplier of 1.0.
 
 The model contains a trained instance of the network, packaged to do the [image classification](https://www.tensorflow.org/hub/common_signatures/images#classification) that the network was trained on. If you merely want to transform images into feature vectors, use [`google/imagenet/mobilenet_v3_small_100_224/feature_vector/5`](https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/feature_vector/5) instead, and save the space occupied by the classification layer.
 
@@ -45,11 +45,21 @@ The checkpoint exported into this model was `v3-small_224_1.0_float/ema/model-38
 
 This model can be used with the `hub.KerasLayer` as follows. It cannot be used with the `hub.Module` API for TensorFlow 1.
 
+### Using TF Hub and HF Hub
 ```
-m = tf.keras.Sequential([
-    hub.KerasLayer("https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/classification/5")
-])
-m.build([None, 224, 224, 3])  # Batch input shape.
+model_path = snapshot_download(repo_id="Dimitre/mobilenet_v3_small")
+model =  KerasLayer(handle=model_path)
+
+img = np.random.rand(1, 224, 224, 3) # (batch_size, height, width, num_channels)
+model(img) # output shape (1, 1001)
+```
+
+### Using [TF Hub fork](https://github.com/dimitreOliveira/hub)
+```
+model = pull_from_hub(repo_id="Dimitre/mobilenet_v3_small")
+
+img = np.random.rand(1, 224, 224, 3) # (batch_size, height, width, num_channels)
+model(img) # output shape (1, 1001)
 ```
 
 The output is a batch of logits vectors. The indices into the logits are the `num_classes` = 1001 classes of the classification from the original training (see above). The mapping from indices to class labels can be found in the file at [download.tensorflow.org/data/ImageNetLabels.txt](https://storage.googleapis.com/download.tensorflow.org/data/ImageNetLabels.txt) (with class 0 for "background", followed by 1000 actual ImageNet classes).
@@ -62,9 +72,21 @@ In principle, consumers of this model can [fine-tune](https://www.tensorflow.org
 
 However, fine-tuning through a large classification might be prone to overfit.
 
-The momentum (a.k.a. decay coefficient) of batch norm's exponential moving averages defaults to 0.99 for this model, in order to accelerate training on small datasets (or with huge batch sizes). Advanced users can set another value (say, 0.997) by loading this model like
+The momentum (a.k.a. decay coefficient) of batch norm's exponential moving averages defaults to 0.99 for this model, in order to accelerate training on small datasets (or with huge batch sizes).
 
+### Using TF Hub and HF Hub
 ```
-hub.KerasLayer("https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/classification/5",
-               trainable=True, arguments=dict(batch_norm_momentum=0.997))
+model_path = snapshot_download(repo_id="Dimitre/mobilenet_v3_small")
+model =  KerasLayer(handle=model_path, trainable=True)
+
+img = np.random.rand(1, 224, 224, 3) # (batch_size, height, width, num_channels)
+model(img) # output shape (1, 1001)
+```
+
+### Using [TF Hub fork](https://github.com/dimitreOliveira/hub)
+```
+model = pull_from_hub(repo_id="Dimitre/mobilenet_v3_small", trainable=True)
+
+img = np.random.rand(1, 224, 224, 3) # (batch_size, height, width, num_channels)
+model(img) # output shape (1, 1001)
 ```