In [ ]:
#@title Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

Guida rapida a Tensorflow 2 per principianti

Note: La nostra comunità di Tensorflow ha tradotto questi documenti. Poichè queste traduzioni sono best-effort, non è garantito che rispecchino in maniera precisa e aggiornata la documentazione ufficiale in inglese. Se avete suggerimenti per migliorare questa traduzione, mandate per favore una pull request al repository Github tensorflow/docs. Per proporsi come volontari alla scrittura o alla review delle traduzioni della comunità contattate la mailing list docs@tensorflow.org.

Questa breve introduzione usa Keras per:

  1. Costruire una rete neurale che classifica immagini.
  2. Addestrare questa rete neurale.
  3. E, infine, valutare l'accuratezza del modello.

Questo è un notebook file. I programmi Python sono eseguiti direttamente nel browser—un ottimo modo per imparare e utilizzare TensorFlow. Per seguire questo tutorial, esegui il file notebook in Google Colab cliccando sul bottone in cima a questa pagina.

  1. All'interno di Colab, connettiti al runtime di Python: In alto a destra della barra dei menu, seleziona CONNECT.
  2. Esegui tutte le celle di codice di notebook: Seleziona Runtime > Run all.

Scarica e installa il package TensorFlow 2. Importa TensorFlow nel tuo codice:


In [ ]:
# Install TensorFlow

import tensorflow as tf

Carica e prepara il dataset MNIST. Converti gli esempi da numeri di tipo integer a floating-point:


In [ ]:
mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

Costruisci il modello tf.keras.Sequential tramite layer. Scegli un metodo di ottimizzazione e una funzione obiettivo per l'addestramento:


In [ ]:
model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(input_shape=(28, 28)),
  tf.keras.layers.Dense(128, activation='relu'),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10, activation='softmax')
])

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

Addestra e valuta il modello:


In [ ]:
model.fit(x_train, y_train, epochs=5)

model.evaluate(x_test,  y_test, verbose=2)

Il classificatore di immagini è ora addestrato per circa il 98% di accuratezza su questo insieme di dati. Per approfondire, leggi il tutorial di TensorFlow.