imbd_inter_text.ipynb

{"cells":[{"cell_type":"markdown","metadata":{"id":"jY1e8NGLKgRo"},"source":["# IMBD : interprétation de texte"]},{"cell_type":"markdown","source":["1. le but de cet exemple est de décider si une critique de film est positive ou négative en fonction de commentaires du film\n","2. le codage des commentaires est expliqué dans le shéma ci-dessous\n","3. le modèle doit trancher si l'avis au global est négatif-0 ou positif-1"],"metadata":{"id":"XjjVk9E1jnNb"}},{"cell_type":"markdown","source":["![codage_critique.png](data:image/png;base64,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)"],"metadata":{"id":"7fO2ePcHyiJV"}},{"cell_type":"markdown","metadata":{"id":"r-68aNpwK0-_"},"source":["## 1. modules à charger"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"sCOoEhnoJvO7"},"outputs":[],"source":["import numpy as np\n","from tensorflow import keras\n","from tensorflow.keras import optimizers\n","from tensorflow.keras.models import Sequential\n","from tensorflow.keras.layers import Dense"]},{"cell_type":"markdown","metadata":{"id":"Sx0rGqTGL90P"},"source":["## 2. Données"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"nc14IE30J-Ko","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1686313119045,"user_tz":-120,"elapsed":5752,"user":{"displayName":"CV Maths","userId":"14309968410644106137"}},"outputId":"8aedea06-01c5-4a7f-e844-ae368d07d15c"},"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz\n","17464789/17464789 [==============================] - 2s 0us/step\n"]}],"source":["from tensorflow.keras.datasets import imdb\n","\n","nb_mots_total = 1000   # On ne garde que les n=1000 mots les plus fréquents \n","(X_train_data, Y_train), (X_test_data, Y_test) = imdb.load_data(num_words=nb_mots_total)"]},{"cell_type":"code","source":["# Afficher une critique et sa note \n","def affiche_texte(num):\n","    index_mots = imdb.get_word_index()\n","    index_mots_inverse = dict([(value, key) for (key, value) in index_mots.items()])\n","    critique_mots = ' '.join([index_mots_inverse.get(i - 3, '??') for i in X_train_data[num]])\n","    print(\"Critique :\\n\", critique_mots)\n","    print(\"Note 0 (négatif) ou 1 (positif) ? :\", Y_train[num])\n","    print(\"Critique (sous forme brute) :\\n\", X_train_data[num])\n","    return\n","\n","affiche_texte(123)   # affichage de la critique numéro 123"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uDxeuL0mFKT6","executionInfo":{"status":"ok","timestamp":1686313125799,"user_tz":-120,"elapsed":864,"user":{"displayName":"CV Maths","userId":"14309968410644106137"}},"outputId":"44c4d763-4cb7-4ab9-b23b-1974efd0a968"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb_word_index.json\n","1641221/1641221 [==============================] - 0s 0us/step\n","Critique :\n"," ?? beautiful and ?? movie ?? ?? great ?? good acting and one of the most ?? movies i have seen in a while i never saw such an interesting setting when i was in ?? my wife liked it so much she ?? me to ?? on and rate it so other would enjoy too\n","Note 0 (négatif) ou 1 (positif) ? : 1\n","Critique (sous forme brute) :\n"," [1, 307, 5, 2, 20, 2, 2, 87, 2, 52, 116, 5, 31, 7, 4, 91, 2, 102, 13, 28, 110, 11, 6, 137, 13, 115, 219, 141, 35, 221, 956, 54, 13, 16, 11, 2, 61, 322, 423, 12, 38, 76, 59, 2, 72, 8, 2, 23, 5, 967, 12, 38, 85, 62, 358, 99]\n"]}]},{"cell_type":"code","source":["# Données sous forme de vecteurs\n","def vectorisation_critiques(X_data):\n","    vecteurs = np.zeros((len(X_data), nb_mots_total))\n","    for i in range(len(X_data)):\n","        for c in X_data[i]:\n","            vecteurs[i,c] = 1.0\n","    return vecteurs\n","\n","X_train = vectorisation_critiques(X_train_data)\n","X_test = vectorisation_critiques(X_test_data)"],"metadata":{"id":"l73q7fujFVCj"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"2sCLejJzMGOY"},"source":["## 3. Réseau"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"CoZOOla3KdSo"},"outputs":[],"source":["modele = Sequential()\n","p = 5\n","modele.add(Dense(p, input_dim=nb_mots_total, activation='relu'))\n","modele.add(Dense(p, activation='relu'))\n","modele.add(Dense(p, activation='relu'))\n","modele.add(Dense(1, activation='sigmoid'))\n","modele.compile(loss='binary_crossentropy', optimizer='sgd', metrics=['accuracy'])"]},{"cell_type":"markdown","metadata":{"id":"C0HlfK7vMO7Q"},"source":["## 4. Entrainement du modèle"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":35698,"status":"ok","timestamp":1686313198983,"user":{"displayName":"CV Maths","userId":"14309968410644106137"},"user_tz":-120},"id":"_nEc672uL6Zf","outputId":"e85979d0-058b-49a4-94dd-b03026f42a65"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/10\n","782/782 [==============================] - 9s 4ms/step - loss: 0.6924 - accuracy: 0.5225\n","Epoch 2/10\n","782/782 [==============================] - 3s 4ms/step - loss: 0.6692 - accuracy: 0.6273\n","Epoch 3/10\n","782/782 [==============================] - 3s 3ms/step - loss: 0.4894 - accuracy: 0.7847\n","Epoch 4/10\n","782/782 [==============================] - 3s 3ms/step - loss: 0.3652 - accuracy: 0.8423\n","Epoch 5/10\n","782/782 [==============================] - 3s 4ms/step - loss: 0.3352 - accuracy: 0.8566\n","Epoch 6/10\n","782/782 [==============================] - 3s 4ms/step - loss: 0.3229 - accuracy: 0.8650\n","Epoch 7/10\n","782/782 [==============================] - 3s 3ms/step - loss: 0.3167 - accuracy: 0.8683\n","Epoch 8/10\n","782/782 [==============================] - 3s 3ms/step - loss: 0.3136 - accuracy: 0.8687\n","Epoch 9/10\n","782/782 [==============================] - 3s 3ms/step - loss: 0.3114 - accuracy: 0.8692\n","Epoch 10/10\n","782/782 [==============================] - 4s 5ms/step - loss: 0.3083 - accuracy: 0.8718\n"]},{"output_type":"execute_result","data":{"text/plain":["<keras.callbacks.History at 0x7f68f223f3a0>"]},"metadata":{},"execution_count":6}],"source":["modele.fit(X_train, Y_train, epochs=10, batch_size=32)"]},{"cell_type":"markdown","metadata":{"id":"FevGG1NLMi3w"},"source":["## 5. Résultats"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":3302,"status":"ok","timestamp":1686313206405,"user":{"displayName":"CV Maths","userId":"14309968410644106137"},"user_tz":-120},"id":"nKvYX2s3MhQh","outputId":"4764a614-8056-48e3-f16b-7b8509ab51e6"},"outputs":[{"output_type":"stream","name":"stdout","text":["782/782 [==============================] - 1s 2ms/step\n"]}],"source":["Y_predict = modele.predict(X_test)"]},{"cell_type":"markdown","source":["## Afficher une critique et sa note "],"metadata":{"id":"JXvvJEuqEflL"}},{"cell_type":"code","source":["def affiche_texte_test(num):\n","    index_mots = imdb.get_word_index()\n","    index_mots_inverse = dict([(value, key) for (key, value) in index_mots.items()])\n","    critique_mots = ' '.join([index_mots_inverse.get(i - 3, '??') for i in X_test_data[num]])\n","    print(\"Critique :\\n\", critique_mots)\n","    print(\"Note attendue 0 (négatif) ou 1 (positif) ? :\", Y_test[num])\n","    print(\"Note prédite 0 (négatif) ou 1 (positif) ? :\", Y_predict[num][0])\n","    return\n","\n","affiche_texte_test(111)   # prédiction pour la critique test numéro 111"],"metadata":{"id":"nbO5L-rEiXGj","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1686313211061,"user_tz":-120,"elapsed":621,"user":{"displayName":"CV Maths","userId":"14309968410644106137"}},"outputId":"baa9466e-2233-4414-f75d-5d3084ee0a1e"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Critique :\n"," ?? i at first thought this little fantasy ?? would be a little entertaining i was wrong br br a good cast ?? ?? as the ?? didn't help it any the story had every ?? possible worst case ?? that could take place in a ?? ?? ?? and none of it could possibly happen br br true the ?? of the ?? could only be ?? with the ?? help of a ?? in the ?? ?? ?? air ?? one but everything they ?? ?? the ?? and the ?? of our country if were to fall into ?? hands is ?? to the ?? ?? seriously not even the ?? can ?? over ?? our ?? ?? the case is only used to ?? ?? in this situation our ?? would have completely ?? the ?? and the whole thing would go ?? the ?? of ?? couldn't happen there would not have been a ?? ?? because the ?? ?? would have been ?? not to ?? ?? a ?? there are just too many ?? ?? to ?? such a thing from ?? br br true film's like ?? ?? and ?? gave some ?? to the ?? of us ?? ?? of the ?? but this film goes too far and fails to ?? my ?? of the ?? and that makes the experience a waste of the ?? time this film is a ??\n","Note attendue 0 (négatif) ou 1 (positif) ? : 0\n","Note prédite 0 (négatif) ou 1 (positif) ? : 0.008406711\n"]}]}],"metadata":{"accelerator":"GPU","colab":{"gpuType":"T4","provenance":[]},"kernelspec":{"display_name":"Python 3 (ipykernel)","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.9.16"}},"nbformat":4,"nbformat_minor":0}