diff --git a/notebooks/Predict_XAI.ipynb b/notebooks/Predict_XAI.ipynb index 347b3d0760c5efe64b1fb3c423220a66abf85a9c..cbc49c8e8d105ce4f4ef511c36ea7fef92f9bc23 100644 --- a/notebooks/Predict_XAI.ipynb +++ b/notebooks/Predict_XAI.ipynb @@ -93,7 +93,7 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 102, "metadata": { "id": "SkErnwgMMbRj" }, @@ -105,6 +105,7 @@ "from transformers import BertTokenizer, BertForSequenceClassification\n", "from transformers_interpret import SequenceClassificationExplainer\n", "\n", + "from tqdm import tqdm\n", "import numpy as np\n", "import torch\n", "from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n", @@ -1268,7 +1269,85 @@ }, { "cell_type": "code", - "execution_count": 59, + "execution_count": 51, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "'Géographie'" + ] + }, + "execution_count": 51, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "cls_explainer.predicted_class_name" + ] + }, + { + "cell_type": "code", + "execution_count": 52, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "<table width: 100%><div style=\"border-top: 1px solid; margin-top: 5px; padding-top: 5px; display: inline-block\"><b>Legend: </b><span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 60%)\"></span> Negative <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 100%)\"></span> Neutral <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(120, 75%, 50%)\"></span> Positive </div><tr><th>True Label</th><th>Predicted Label</th><th>Attribution Label</th><th>Attribution Score</th><th>Word Importance</th><tr><td><text style=\"padding-right:2em\"><b>15</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie (1.00)</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie</b></text></td><td><text style=\"padding-right:2em\"><b>6.80</b></text></td><td><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [CLS] </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> L </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##Y </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ON </font></mark><mark style=\"background-color: hsl(120, 75%, 91%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ( </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> G </font></mark><mark style=\"background-color: hsl(120, 75%, 89%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##éo </font></mark><mark style=\"background-color: hsl(120, 75%, 92%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gr </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(0, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ) </font></mark><mark style=\"background-color: hsl(120, 75%, 90%); opacity:1.0; line-height:1.75\"><font color=\"black\"> grande </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> riche </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> belle </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ancienne </font></mark><mark style=\"background-color: hsl(0, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> celebre </font></mark><mark style=\"background-color: hsl(120, 75%, 77%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ville </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> France </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> plus </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> con </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##sid </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##érable </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> royaume </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> après </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Paris </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 86%); opacity:1.0; line-height:1.75\"><font color=\"black\"> capitale </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##noi </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Elle </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> nomme </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> en </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> latin </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##dun </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gus </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##iano </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##rum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Cel </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##tar </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> c </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Vo </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ye </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##z </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fut </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fondée </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> an </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rome </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> 712 </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> quarante </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> - </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> un </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ans </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avant </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ere </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> chrétienne </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> par </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lucius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Mu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##nati </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Plan </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##cus </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> é </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##toi </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##t </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> consul </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avec </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> AE </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##mil </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Le </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pid </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Il </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> b </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ât </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##it </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> sur </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> S </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ôn </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> au </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> lieu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> où </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> cette </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> rivier </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> jet </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##te </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> dans </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> le </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rhône </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> il </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> peu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pla </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> des </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> citoyens </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> romain </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> a </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [SEP] </font></mark></td><tr></table>" + ], + "text/plain": [ + "<IPython.core.display.HTML object>" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "text/html": [ + "<table width: 100%><div style=\"border-top: 1px solid; margin-top: 5px; padding-top: 5px; display: inline-block\"><b>Legend: </b><span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 60%)\"></span> Negative <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 100%)\"></span> Neutral <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(120, 75%, 50%)\"></span> Positive </div><tr><th>True Label</th><th>Predicted Label</th><th>Attribution Label</th><th>Attribution Score</th><th>Word Importance</th><tr><td><text style=\"padding-right:2em\"><b>15</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie (1.00)</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie</b></text></td><td><text style=\"padding-right:2em\"><b>6.80</b></text></td><td><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [CLS] </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> L </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##Y </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ON </font></mark><mark style=\"background-color: hsl(120, 75%, 91%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ( </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> G </font></mark><mark style=\"background-color: hsl(120, 75%, 89%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##éo </font></mark><mark style=\"background-color: hsl(120, 75%, 92%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gr </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(0, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ) </font></mark><mark style=\"background-color: hsl(120, 75%, 90%); opacity:1.0; line-height:1.75\"><font color=\"black\"> grande </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> riche </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> belle </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ancienne </font></mark><mark style=\"background-color: hsl(0, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> celebre </font></mark><mark style=\"background-color: hsl(120, 75%, 77%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ville </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> France </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> plus </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> con </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##sid </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##érable </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> royaume </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> après </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Paris </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 86%); opacity:1.0; line-height:1.75\"><font color=\"black\"> capitale </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##noi </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Elle </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> nomme </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> en </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> latin </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##dun </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gus </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##iano </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##rum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Cel </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##tar </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> c </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Vo </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ye </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##z </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fut </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fondée </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> an </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rome </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> 712 </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> quarante </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> - </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> un </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ans </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avant </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ere </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> chrétienne </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> par </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lucius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Mu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##nati </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Plan </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##cus </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> é </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##toi </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##t </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> consul </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avec </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> AE </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##mil </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Le </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pid </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Il </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> b </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ât </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##it </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> sur </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> S </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ôn </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> au </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> lieu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> où </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> cette </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> rivier </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> jet </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##te </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> dans </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> le </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rhône </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> il </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> peu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pla </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> des </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> citoyens </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> romain </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> a </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [SEP] </font></mark></td><tr></table>" + ], + "text/plain": [ + "<IPython.core.display.HTML object>" + ] + }, + "execution_count": 52, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "cls_explainer.visualize()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "* récupérer les mots positifs par domaine (EDdA et LGE)\n", + "* faire des nuages de mots et comparer les plus fréquents entre EDdA et LGE (corpus parallèle)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "pour chaque domaine\n", + "pour chaque token : moyenne (somme ?) des scores ? ne prendre que les scores au dessus de 0.1 ?" + ] + }, + { + "cell_type": "code", + "execution_count": 63, "metadata": {}, "outputs": [ { @@ -1436,7 +1515,7 @@ " (')', -0.13544847084394057)]" ] }, - "execution_count": 59, + "execution_count": 63, "metadata": {}, "output_type": "execute_result" } @@ -1448,77 +1527,1502 @@ }, { "cell_type": "code", - "execution_count": 51, + "execution_count": 83, + "metadata": {}, + "outputs": [], + "source": [ + "d = {}" + ] + }, + { + "cell_type": "code", + "execution_count": 84, "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "'Géographie'" + "{'ville': 0.478071716663547,\n", + " 'capitale': 0.2983988672217172,\n", + " 'royaume': 0.24665610131446675,\n", + " 'G': 0.24402357535335403,\n", + " '##éo': 0.23393328870446992,\n", + " 'grande': 0.21832893139528123,\n", + " ',': -0.011642202072501788,\n", + " '##gr': 0.1695800465119405,\n", + " '##Y': 0.1456759996705617,\n", + " '##ON': 0.14307146561933012,\n", + " '.': 0.029180628239546275,\n", + " 'France': 0.13688799086603975,\n", + " 'latin': 0.13303588704102381,\n", + " 'chrétienne': 0.1311835388990743,\n", + " 'lieu': 0.13020947076813982,\n", + " 'ancienne': 0.12958979621300132,\n", + " 'celebre': 0.12947489123965564,\n", + " 'rivier': 0.11169096058453537,\n", + " 'nomme': 0.10000471924693329,\n", + " '(': 0.09932002907423143,\n", + " 'la': -0.006879341345145134,\n", + " 'riche': 0.07913704700022943,\n", + " 'du': 0.018572791985543135,\n", + " 'Rhône': 0.07226400802922804,\n", + " 'Elle': 0.0693630173969722,\n", + " \"'\": 0.05723696778164145,\n", + " 'S': 0.06575101714951456,\n", + " 'plus': 0.05840061099213507,\n", + " 'qui': 0.03778469886529954,\n", + " '712': 0.054693452829347115,\n", + " '##érable': 0.05297839086419718,\n", + " 'fondée': 0.04704211890403151,\n", + " 'l': 0.0451974674122074,\n", + " 'fut': 0.0438798787047486,\n", + " '##e': 0.010537012868472688,\n", + " 'b': 0.042870227388534604,\n", + " 'sur': 0.04260004363332922,\n", + " 'se': 0.0058271154715995995,\n", + " 'peu': 0.040981027718879084,\n", + " 'c': 0.0403850871592572,\n", + " 'Lyon': 0.022692171217471906,\n", + " '##um': 0.03484266276127894,\n", + " '##ye': 0.03940461731845493,\n", + " '##rum': 0.03820084664850618,\n", + " '##m': 0.026521004509341334,\n", + " 'Se': 0.037503453809376,\n", + " 'cette': 0.03737564027887762,\n", + " 'Rome': 0.03688032185991681,\n", + " '##umu': 0.03456580806237662,\n", + " 'de': 0.021013220187771894,\n", + " 'Paris': 0.03310146903416289,\n", + " '##gus': 0.03294045015997047,\n", + " '##sid': 0.03255043778254519,\n", + " 'jet': 0.031575857152632385,\n", + " 'un': 0.03117251985912735,\n", + " 'le': 0.03105610182850656,\n", + " 'Lu': 0.008645628419735481,\n", + " 'Plan': 0.030820184404097863,\n", + " 'en': 0.03010674205624715,\n", + " '##z': 0.029871874749211054,\n", + " '##mil': 0.02963200210194755,\n", + " '##unum': 0.025898349689579492,\n", + " 'Vo': 0.02935262786796574,\n", + " '##noi': 0.02909189419875202,\n", + " 'des': 0.028996120278423045,\n", + " 'quarante': 0.028231791558966633,\n", + " '##s': 0.01484737615013025,\n", + " 'con': 0.026165582559808873,\n", + " '##dun': 0.025699022336446258,\n", + " '##tar': 0.025188870477124894,\n", + " 'é': 0.024089543382319098,\n", + " 'an': 0.023585319400848195,\n", + " '##gu': 0.023084632572130535,\n", + " 'Il': 0.021736540370470812,\n", + " '##us': 0.015200983089939281,\n", + " '##te': 0.020236291895152022,\n", + " 'avec': 0.019039309232488966,\n", + " 'dans': 0.01852231748257226,\n", + " '##cus': 0.01828726599412002,\n", + " '##pid': 0.01787476167297771,\n", + " 'après': 0.01785470962170739,\n", + " 'avant': 0.017626577836139475,\n", + " '##nati': 0.01752347206998558,\n", + " '##iano': 0.017089445343453365,\n", + " 'Le': 0.016166723086828174,\n", + " '##toi': 0.01483008688193065,\n", + " 'ans': 0.014472180695321534,\n", + " 'Cel': 0.014027086848242715,\n", + " 'où': 0.013157964330803138,\n", + " '##ât': 0.011332787999157318,\n", + " '##ôn': 0.011241165099203603,\n", + " '##it': 0.009763016011555254,\n", + " 'ere': 0.007590037219544403,\n", + " 'L': 0.007399733805079844,\n", + " 'par': 0.007011176299182855,\n", + " '##gd': -0.00442376201350928,\n", + " 'Mu': 0.006340399133187405,\n", + " 'consul': 0.006312700914285012,\n", + " 'il': 0.00609352197030786,\n", + " '##pla': 0.005560350755837545,\n", + " 'Lucius': 0.0044462351021057325,\n", + " '##t': 0.0036622619849812073,\n", + " '[CLS]': 0.0,\n", + " '[SEP]': 0.0,\n", + " '-': -0.0003846539976056082,\n", + " '##ius': -0.00048531039895657175,\n", + " 'au': -0.0018304190363696647,\n", + " '&': -0.08418116246612357,\n", + " 'romain': -0.008083189911088765,\n", + " 'AE': -0.010990138859793724,\n", + " 'a': -0.022083265525204197,\n", + " 'belle': -0.029909244412604778,\n", + " 'citoyens': -0.03278504989463669,\n", + " ')': -0.13544847084394057}" ] }, - "execution_count": 51, + "execution_count": 84, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "cls_explainer.predicted_class_name" + "d_tmp = dict(word_attributions)\n", + "d_tmp" ] }, { "cell_type": "code", - "execution_count": 52, + "execution_count": 97, "metadata": {}, "outputs": [ { "data": { - "text/html": [ - "<table width: 100%><div style=\"border-top: 1px solid; margin-top: 5px; padding-top: 5px; display: inline-block\"><b>Legend: </b><span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 60%)\"></span> Negative <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 100%)\"></span> Neutral <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(120, 75%, 50%)\"></span> Positive </div><tr><th>True Label</th><th>Predicted Label</th><th>Attribution Label</th><th>Attribution Score</th><th>Word Importance</th><tr><td><text style=\"padding-right:2em\"><b>15</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie (1.00)</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie</b></text></td><td><text style=\"padding-right:2em\"><b>6.80</b></text></td><td><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [CLS] </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> L </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##Y </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ON </font></mark><mark style=\"background-color: hsl(120, 75%, 91%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ( </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> G </font></mark><mark style=\"background-color: hsl(120, 75%, 89%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##éo </font></mark><mark style=\"background-color: hsl(120, 75%, 92%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gr </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(0, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ) </font></mark><mark style=\"background-color: hsl(120, 75%, 90%); opacity:1.0; line-height:1.75\"><font color=\"black\"> grande </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> riche </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> belle </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ancienne </font></mark><mark style=\"background-color: hsl(0, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> celebre </font></mark><mark style=\"background-color: hsl(120, 75%, 77%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ville </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> France </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> plus </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> con </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##sid </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##érable </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> royaume </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> après </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Paris </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 86%); opacity:1.0; line-height:1.75\"><font color=\"black\"> capitale </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##noi </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Elle </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> nomme </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> en </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> latin </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##dun </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gus </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##iano </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##rum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Cel </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##tar </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> c </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Vo </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ye </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##z </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fut </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fondée </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> an </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rome </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> 712 </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> quarante </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> - </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> un </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ans </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avant </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ere </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> chrétienne </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> par </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lucius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Mu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##nati </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Plan </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##cus </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> é </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##toi </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##t </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> consul </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avec </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> AE </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##mil </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Le </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pid </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Il </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> b </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ât </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##it </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> sur </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> S </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ôn </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> au </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> lieu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> où </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> cette </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> rivier </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> jet </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##te </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> dans </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> le </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rhône </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> il </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> peu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pla </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> des </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> citoyens </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> romain </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> a </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [SEP] </font></mark></td><tr></table>" - ], "text/plain": [ - "<IPython.core.display.HTML object>" + "array([nan, 'Grammaire', 'Médailles', 'Histoire', 'Caractères',\n", + " 'Philosophie', 'Chimie', 'Médecine - Chirurgie', 'Commerce',\n", + " 'Géographie', 'Mathématiques', 'Histoire naturelle',\n", + " 'Architecture', 'Blason', 'Agriculture - Economie rustique',\n", + " 'Chasse', 'Métiers', 'Anatomie', 'Droit - Jurisprudence',\n", + " 'Antiquité', 'Marine', 'Militaire (Art) - Guerre - Arme',\n", + " 'Economie domestique', 'Maréchage - Manège', 'Jeu', 'Monnaie',\n", + " 'Physique - [Sciences physico-mathématiques]', 'Religion', 'Pêche',\n", + " 'Pharmacie', 'Musique', 'Beaux-arts', 'Spectacle', 'Politique',\n", + " 'Mesure', 'Belles-lettres - Poésie', 'Superstition',\n", + " 'Arts et métiers', 'Minéralogie'], dtype=object)" ] }, + "execution_count": 97, "metadata": {}, - "output_type": "display_data" - }, + "output_type": "execute_result" + } + ], + "source": [ + "df_EDdA.enccre_class.unique()" + ] + }, + { + "cell_type": "code", + "execution_count": 105, + "metadata": {}, + "outputs": [], + "source": [ + "def get_dict_xai(word_attributions):\n", + "\n", + " word_attributions.sort(key=lambda a: a[1], reverse = True)\n", + " return dict(word_attributions)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 108, + "metadata": {}, + "outputs": [], + "source": [ + "df = df_EDdA.copy()" + ] + }, + { + "cell_type": "code", + "execution_count": 109, + "metadata": {}, + "outputs": [], + "source": [ + "df = df.sample(100)" + ] + }, + { + "cell_type": "code", + "execution_count": 111, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "2it [00:05, 2.71s/it]\n", + "3it [00:13, 4.63s/it]\n", + "3it [00:10, 3.66s/it]\n", + "2it [00:05, 2.57s/it]\n", + "1it [00:01, 1.19s/it]\n", + "1it [00:02, 2.46s/it]\n", + "1it [00:01, 1.73s/it]\n", + "1it [00:03, 3.13s/it]\n", + "2it [00:09, 4.66s/it]\n", + "11it [00:21, 1.99s/it]\n", + "1it [00:01, 1.41s/it]\n", + "16it [00:47, 2.97s/it]\n", + "2it [00:06, 3.33s/it]\n", + "10it [00:40, 4.09s/it]\n", + "3it [00:06, 2.25s/it]\n", + "1it [00:01, 1.14s/it]\n", + "1it [00:01, 1.60s/it]\n", + "2it [00:06, 3.00s/it]\n", + "1it [00:01, 1.60s/it]\n", + "1it [00:02, 2.46s/it]\n", + "1it [00:00, 1.30it/s]\n", + "4it [00:15, 3.97s/it]\n", + "4it [00:13, 3.41s/it]\n", + "2it [00:07, 3.59s/it]\n", + "1it [00:04, 4.69s/it]\n", + "1it [00:00, 1.24it/s]\n" + ] + } + ], + "source": [ + "d = {}\n", + "# boucle sur les domaines\n", + "for group_name, df_group in df.groupby('enccre_class'):\n", + " d[group_name] = {}\n", + " # boucle sur les articles\n", + " for index, row in tqdm(df_group.iterrows()):\n", + " \n", + " word_attributions = cls_explainer(row['content'] if len(row['content']) < 512 else row['content'][:512])\n", + "\n", + " d_tmp = get_dict_xai(word_attributions)\n", + "\n", + " for k, v in d_tmp.items():\n", + " if k in d[group_name]:\n", + " d[group_name][k] = (d[group_name][k] + v) / 2\n", + " else:\n", + " d[group_name][k] = v\n" + ] + }, + { + "cell_type": "code", + "execution_count": 112, + "metadata": {}, + "outputs": [ { "data": { - "text/html": [ - "<table width: 100%><div style=\"border-top: 1px solid; margin-top: 5px; padding-top: 5px; display: inline-block\"><b>Legend: </b><span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 60%)\"></span> Negative <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(0, 75%, 100%)\"></span> Neutral <span style=\"display: inline-block; width: 10px; height: 10px; border: 1px solid; background-color: hsl(120, 75%, 50%)\"></span> Positive </div><tr><th>True Label</th><th>Predicted Label</th><th>Attribution Label</th><th>Attribution Score</th><th>Word Importance</th><tr><td><text style=\"padding-right:2em\"><b>15</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie (1.00)</b></text></td><td><text style=\"padding-right:2em\"><b>Géographie</b></text></td><td><text style=\"padding-right:2em\"><b>6.80</b></text></td><td><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [CLS] </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> L </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##Y </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ON </font></mark><mark style=\"background-color: hsl(120, 75%, 91%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ( </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> G </font></mark><mark style=\"background-color: hsl(120, 75%, 89%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##éo </font></mark><mark style=\"background-color: hsl(120, 75%, 92%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gr </font></mark><mark style=\"background-color: hsl(120, 75%, 93%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(0, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ) </font></mark><mark style=\"background-color: hsl(120, 75%, 90%); opacity:1.0; line-height:1.75\"><font color=\"black\"> grande </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> riche </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> belle </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ancienne </font></mark><mark style=\"background-color: hsl(0, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> celebre </font></mark><mark style=\"background-color: hsl(120, 75%, 77%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ville </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> France </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> plus </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> con </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##sid </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##érable </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 88%); opacity:1.0; line-height:1.75\"><font color=\"black\"> royaume </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> après </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Paris </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 96%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 86%); opacity:1.0; line-height:1.75\"><font color=\"black\"> capitale </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> du </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##noi </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Elle </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> nomme </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> en </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> latin </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##dun </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gus </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##iano </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##rum </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##umu </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##m </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Cel </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##tar </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##um </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> c </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Vo </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ye </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##z </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lu </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##gd </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##unum </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lyon </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fut </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> fondée </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> an </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> de </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rome </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> 712 </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> quarante </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> - </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> un </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ans </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avant </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> l </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ' </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ere </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> chrétienne </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> par </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Lucius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Mu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##nati </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Plan </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##cus </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> é </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##toi </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##t </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> consul </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> avec </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> AE </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##mil </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ius </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Le </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pid </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##us </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> . </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Il </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> b </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ât </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##it </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> sur </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> S </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##ôn </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> au </font></mark><mark style=\"background-color: hsl(120, 75%, 94%); opacity:1.0; line-height:1.75\"><font color=\"black\"> lieu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> où </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> cette </font></mark><mark style=\"background-color: hsl(120, 75%, 95%); opacity:1.0; line-height:1.75\"><font color=\"black\"> rivier </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##e </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> se </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> jet </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##te </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> dans </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> le </font></mark><mark style=\"background-color: hsl(120, 75%, 97%); opacity:1.0; line-height:1.75\"><font color=\"black\"> Rhône </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> , </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> & </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> il </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> la </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> peu </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##pla </font></mark><mark style=\"background-color: hsl(120, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> des </font></mark><mark style=\"background-color: hsl(0, 75%, 99%); opacity:1.0; line-height:1.75\"><font color=\"black\"> citoyens </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> romain </font></mark><mark style=\"background-color: hsl(120, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> ##s </font></mark><mark style=\"background-color: hsl(120, 75%, 98%); opacity:1.0; line-height:1.75\"><font color=\"black\"> qui </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> a </font></mark><mark style=\"background-color: hsl(0, 75%, 100%); opacity:1.0; line-height:1.75\"><font color=\"black\"> [SEP] </font></mark></td><tr></table>" - ], "text/plain": [ - "<IPython.core.display.HTML object>" + "{'ville': 0.468205314323041,\n", + " 'fleuve': 0.2423772411435995,\n", + " 'G': 0.23135720218825717,\n", + " ',': 0.0858046931906998,\n", + " 'nommée': 0.19227297918896974,\n", + " '##éo': 0.18973738733606565,\n", + " '##GE': 0.16634948535987606,\n", + " '##A': 0.15877744098706795,\n", + " '##gr': 0.16202194372565118,\n", + " '##TI': 0.14030883450630402,\n", + " '.': 0.054787505866608105,\n", + " '##c': 0.0948633585054219,\n", + " '(': 0.0413040601926874,\n", + " 'depuis': 0.1051878315897289,\n", + " 'nom': 0.18136152655341967,\n", + " 'se': 0.08786898558747075,\n", + " '##ce': 0.09317071188401634,\n", + " '##ia': 0.031040617643215574,\n", + " 'Allemands': 0.0787047715309721,\n", + " 'Ce': 0.06555545762824336,\n", + " 'est': 0.074775644174404,\n", + " 'romaine': 0.07262679757314543,\n", + " 'roi': 0.0716659059875927,\n", + " 'Da': 0.0714045742494807,\n", + " '##é': 0.02500576025356923,\n", + " 'la': 0.12298598711273313,\n", + " 'mais': 0.05591848646995751,\n", + " 'T': 0.0541465558097154,\n", + " 'Sar': 0.01724425985017986,\n", + " 'La': 0.051446794552354624,\n", + " 'ensuite': 0.051200901865576555,\n", + " 'jet': 0.0829284321696537,\n", + " 'J': 0.07546336834323314,\n", + " 'république': 0.04943046233577988,\n", + " 'ce': 0.057519719498181784,\n", + " 'ou': 0.08969930275580645,\n", + " 'an': 0.041146473695852476,\n", + " 'moderne': 0.04177711856360317,\n", + " '##ché': 0.040774048966399426,\n", + " '##toi': 0.03342140264875253,\n", + " 'le': 0.05642199866799782,\n", + " '##ès': 0.11698138135400395,\n", + " '##jana': 0.03818439831225535,\n", + " 'les': 0.0757883643580965,\n", + " 'Le': 0.03715890519752807,\n", + " '##on': 0.025900328149956048,\n", + " '##a': 0.04319994977377055,\n", + " 'Sir': 0.03642712798689307,\n", + " 'Tra': 0.030808153025433653,\n", + " 'avo': 0.032212276900181,\n", + " 'de': 0.06770303542758496,\n", + " '##z': 0.03365499782170345,\n", + " \"'\": 0.061565814719079584,\n", + " 'Ar': 0.032402384389004744,\n", + " 'un': 0.03322588462984368,\n", + " 'sa': 0.032102001051091615,\n", + " 'Hong': 0.03178504105089397,\n", + " '##us': 0.03713196635047464,\n", + " '##bal': 0.03142378925271683,\n", + " '##thus': 0.029706139074928077,\n", + " '##hab': 0.02854877671553812,\n", + " '##rig': 0.028440723473813656,\n", + " '##lp': 0.028348485933541584,\n", + " 'D': 0.021142184892655142,\n", + " '##ssent': 0.026560785640090002,\n", + " 'ar': 0.02655491286531302,\n", + " '##s': 0.047794961185982546,\n", + " 'dont': 0.02121929633019263,\n", + " 'dit': 0.018481328341026063,\n", + " '##cus': 0.024522117453619832,\n", + " '##ziu': 0.023636246682512362,\n", + " 'celui': 0.023612546696296227,\n", + " 'Sam': 0.023408934631155117,\n", + " 'd': 0.05567218618641231,\n", + " 'dans': 0.10022254602904804,\n", + " '##gent': 0.01748654204829846,\n", + " '##noi': 0.02239494074557525,\n", + " '##ros': 0.021349833073345013,\n", + " 'Ist': 0.020901809554305005,\n", + " 'con': 0.025748442770215568,\n", + " '##x': 0.02002047701965849,\n", + " 'R': 0.030949942486234276,\n", + " 'app': 0.018555258608997136,\n", + " '##jano': 0.01808523851267019,\n", + " 'Dion': 0.015754758531027112,\n", + " '##zog': 0.014949231310082899,\n", + " 'U': 0.014654267300384672,\n", + " 'à ': 0.08298650083974059,\n", + " '##oe': 0.013082684704750751,\n", + " 'sous': -0.0002069181948599748,\n", + " 'selon': -0.020082683114963043,\n", + " '##mi': 0.008906287279368003,\n", + " 'Cassius': 0.008545442448074718,\n", + " 'par': 0.061077180857349866,\n", + " '-': 0.06475762097649433,\n", + " '##t': 0.0048467472813908796,\n", + " '##zet': 0.0063275438230979185,\n", + " '##uy': 0.006086328817223191,\n", + " '##el': 0.0060696145131198575,\n", + " 'ca': 0.0056714294799304005,\n", + " '##roi': 0.005468227545142956,\n", + " 'trés': 0.002824228104975018,\n", + " '##oit': 0.001711933310961672,\n", + " '##bu': 0.0016690113836018768,\n", + " '[CLS]': 0.0,\n", + " '[SEP]': 0.0,\n", + " 'ses': 0.049645402358164066,\n", + " '##ors': -0.0017502978171956062,\n", + " 'in': -0.003095317243436783,\n", + " 'que': 0.009635328008189985,\n", + " 'creu': -0.007475521658293897,\n", + " '&': -0.08000801041182604,\n", + " '##it': 0.0015963173159964771,\n", + " ';': 0.007984840521573309,\n", + " ')': -0.06872271549404119,\n", + " 'sentiment': -0.048171146031505886,\n", + " 'SAR': -0.07143494474675065,\n", + " 'latin': 0.3960803982086443,\n", + " 'vallée': 0.24144223688497182,\n", + " 'grec': 0.24154434016104503,\n", + " 'distingue': 0.20225033977383755,\n", + " 'en': 0.026650124208107327,\n", + " '##g': 0.11139228310919369,\n", + " 'une': 0.0705267038169397,\n", + " 'exemple': 0.02878307063738609,\n", + " 'torre': 0.010422364080421317,\n", + " 'fonte': 0.07176335280650341,\n", + " 'mod': 0.05471099375321398,\n", + " 'neige': 0.06716279133677414,\n", + " '##nt': -0.010146726047092878,\n", + " 'des': 0.10395928916105898,\n", + " '##al': 0.021472918740770543,\n", + " 'l': 0.07719865579756106,\n", + " 'toujours': 0.05654656086498567,\n", + " 'grandes': 0.0500708177588014,\n", + " '##ns': 0.010472519252392662,\n", + " 'h': -0.007529062495164878,\n", + " 'aussi': 0.04182030961442956,\n", + " 'du': 0.02917787047349929,\n", + " '##rro': 0.041311241708329303,\n", + " '##u': 0.018374918227908328,\n", + " 'L': 0.03595023525894895,\n", + " 'bien': 0.0283913554226912,\n", + " 'autre': 0.057531150082004046,\n", + " 'plu': 0.03005047616189665,\n", + " '##ule': 0.0166859230453766,\n", + " 'On': 0.008215765462606122,\n", + " '##éra': 0.009599163212087692,\n", + " 'qu': 0.021413621005676176,\n", + " '##bre': 0.01200562189665623,\n", + " 'terme': 0.02610804943109468,\n", + " '##ima': 0.03813331812513627,\n", + " 'E': 0.02353268737063088,\n", + " 'é': 0.031069540958438446,\n", + " 'n': 0.022605076586881327,\n", + " 'puisque': 0.022114351087644566,\n", + " 'cela': 0.021728661127459404,\n", + " 'on': 0.031637879100526636,\n", + " 'ne': 0.023481641713517833,\n", + " '##euse': 0.019031264633545753,\n", + " 'souvent': 0.01412861492800689,\n", + " 'fort': 0.04442544031371985,\n", + " 'danger': 0.013811198066105809,\n", + " 'signifie': 0.013575249323296856,\n", + " 'après': 0.01941062923428665,\n", + " 'met': 0.011327151448637373,\n", + " '##ies': 0.01091580579338499,\n", + " '##que': 0.04353082158018988,\n", + " '##qui': 0.009204343050603915,\n", + " 'che': 0.005705843958191305,\n", + " 'pour': 0.0021363265380875763,\n", + " '##crit': 0.004866795460923155,\n", + " 'co': 0.006828431290704029,\n", + " 'pas': 0.006987930951319245,\n", + " '##vo': -0.007773218105616921,\n", + " '##re': 0.015364885611094212,\n", + " '##ure': -0.029633249072781927,\n", + " 'tem': -0.03507853370176394,\n", + " 'nach': -0.059647272844616725,\n", + " 'to': -0.043587086241707595,\n", + " 'Torre': -0.07370789174796667,\n", + " 'Comme': -0.08102775997694327,\n", + " 'lac': 0.2627689961654854,\n", + " 'duché': 0.15878063729951344,\n", + " 'rivier': 0.1835301313877702,\n", + " 'Novo': 0.12531683029473817,\n", + " '##US': 0.0822458506910208,\n", + " 'empire': 0.12043364298838244,\n", + " 'endroit': 0.12779042183921233,\n", + " 'Russie': 0.10463858590067449,\n", + " 'sur': 0.04915812925999788,\n", + " '##IA': 0.08076525751909835,\n", + " '##SA': 0.02594775540828238,\n", + " '##gor': 0.08311355621144252,\n", + " '##O': 0.12780809591110026,\n", + " 'STAR': 0.037760347726659545,\n", + " 'où': 0.05080575319876235,\n", + " 'Il': 0.0284042208786913,\n", + " '##e': 0.03364232916886872,\n", + " '##vat': 0.05917044284419409,\n", + " '##n': 0.04476191426913992,\n", + " 'Lo': 0.05322826492644412,\n", + " '##te': 0.05775689209126553,\n", + " '##men': 0.025864880430748294,\n", + " '##od': 0.022641329488828195,\n", + " 'province': 0.3138687127885894,\n", + " 'lieu': 0.288944483446071,\n", + " 'septentrional': 0.21736956560941825,\n", + " 'bourg': 0.205192644118189,\n", + " '##er': 0.20470878331452866,\n", + " 'Amérique': 0.2577046079084956,\n", + " 'Frontera': 0.19152877305938218,\n", + " 'deux': 0.07898332743727765,\n", + " 'Espagne': 0.09686731766445605,\n", + " 'Gal': 0.09024113726333705,\n", + " 'nouvelle': 0.0828520285184145,\n", + " 'Guadalajara': 0.0857800140375107,\n", + " '##ice': 0.06692411701657223,\n", + " '##uat': 0.0508369206389474,\n", + " 'X': 0.04064410422404537,\n", + " '##la': 0.028711752575946197,\n", + " '##es': 0.022001482046772214,\n", + " '30': 0.05213470124947088,\n", + " '##ades': -0.01685895398059224,\n", + " 'audience': -0.045768989771362824,\n", + " 'petite': 0.21919648886046295,\n", + " 'royaume': 0.2573068212004148,\n", + " 'Naples': 0.23150278005781894,\n", + " '##Z': 0.1830676183440645,\n", + " 'Italie': 0.1575315512120412,\n", + " '##J': 0.1573522788542329,\n", + " 'terre': 0.08085795705733743,\n", + " '32': 0.12352770141613803,\n", + " '10': 0.07431163055453546,\n", + " 'au': 0.034992091061273066,\n", + " 'Long': 0.06868800126003936,\n", + " '41': 0.08953747205002806,\n", + " 'Labour': 0.07756492814166974,\n", + " 'CA': 0.05046508890359498,\n", + " 'lat': 0.05319105448357667,\n", + " 'Macédoine': 0.3205630652048948,\n", + " '##RI': 0.16460252115303084,\n", + " 'partie': 0.15845359313084062,\n", + " 'lieux': 0.15247473369694478,\n", + " 'habita': 0.13771036963333821,\n", + " '##I': 0.141653889530405,\n", + " 'nord': 0.242739838786995,\n", + " '##E': 0.12333498427020301,\n", + " 'orientale': 0.12299072447402201,\n", + " 'golf': 0.11792822736721131,\n", + " 'A': 0.06479200165315122,\n", + " 'commun': 0.08994387122081324,\n", + " 'voir': 0.08847236372959857,\n", + " '##rma': 0.07365816308437048,\n", + " '##É': 0.07159170704085929,\n", + " '##p': 0.07104594769838767,\n", + " 'P': 0.03272049148432894,\n", + " 'comme': 0.06549376453090695,\n", + " '##mée': 0.0646449194613384,\n", + " '##ïque': 0.061645981622394784,\n", + " '##én': 0.057367966434507346,\n", + " 'Ali': 0.055659756416669474,\n", + " 'C': 0.04306076808452276,\n", + " 'nomme': 0.05401793892863615,\n", + " 'Lu': 0.05137836438835051,\n", + " '##ii': 0.049587268182288424,\n", + " '##olo': 0.04880284359551007,\n", + " 'midi': 0.07113877841272949,\n", + " 'il': 0.0027654637201536968,\n", + " 'va': 0.04337036149864036,\n", + " 'Pier': 0.022936152887815472,\n", + " 'Pt': 0.04241563924208806,\n", + " '##rab': 0.0423110927679243,\n", + " '##rée': 0.042257144545369996,\n", + " '##iot': 0.04203676192154972,\n", + " 'The': 0.04191806043439661,\n", + " 'St': 0.04123715089841821,\n", + " 'Per': 0.03985223606785319,\n", + " '##j': 0.03926027199963966,\n", + " '##ée': 0.039227260949307896,\n", + " 'côté': 0.023297246836922872,\n", + " 'cont': 0.03657382982964496,\n", + " '##ae': 0.03185063483930072,\n", + " '##ac': 0.031224938425030026,\n", + " 'ex': 0.020417873274345934,\n", + " 'différentes': 0.030718941715107615,\n", + " '2°': 0.030522230876155416,\n", + " 'VII': 0.02928677630182494,\n", + " '##dia': 0.026233515028085855,\n", + " '##cer': 0.023491982765047097,\n", + " '##xiu': 0.022805612841692573,\n", + " 'a': -0.0012715805160076403,\n", + " 'commence': 0.014379318563949204,\n", + " '##pt': 0.013195773759531383,\n", + " '##ie': 0.017838742894118582,\n", + " 'cha': 0.0067258655884391136,\n", + " 'fine': 0.006245003686740857,\n", + " '##mon': 0.002094276853030689,\n", + " '##éir': 0.0059480290294971795,\n", + " 'III': 0.00283531432530362,\n", + " 'termine': 0.0013832185393301071,\n", + " 'x': 0.0008492312750708209,\n", + " 'donne': -0.0033814500703101104,\n", + " 'liv': -0.01661384524933283,\n", + " 'born': -0.041898895862637786,\n", + " '1°': -0.026381150057445416,\n", + " 'Asie': 0.25951725852047475,\n", + " 'Orient': 0.26844923609680227,\n", + " 'Inde': 0.2133903020411604,\n", + " 'mer': 0.26868989143550776,\n", + " 'desert': 0.15074697961858496,\n", + " 'petit': 0.1376690268440502,\n", + " 'Cam': 0.12958328223340024,\n", + " '##MP': 0.10974714271316263,\n", + " '##chin': 0.0894472877928192,\n", + " 'Co': 0.07783219273122367,\n", + " '##chine': 0.07614065102466291,\n", + " '##bo': 0.06026845472373662,\n", + " 'ori': 0.05756818718647703,\n", + " '##cci': 0.05157091008708276,\n", + " '##ge': 0.04400356682753303,\n", + " 'o': 0.04070060977171756,\n", + " '##ent': 0.023156820087233715,\n", + " 'CIA': 0.020303354874778645,\n", + " '##dent': 0.00517282162953671,\n", + " 'capitale': 0.36838594655129425,\n", + " 'comté': 0.29220668870626426,\n", + " 'milles': 0.18218276737467964,\n", + " 'Irlande': 0.17055004990277814,\n", + " 'Dublin': 0.14585461747071327,\n", + " 'forte': 0.09134790101615797,\n", + " 'M': 0.04357113158723378,\n", + " '28': 0.11327727914116813,\n", + " '##AR': 0.0700853626124515,\n", + " '40': 0.10679936658542148,\n", + " 'West': 0.08993142493114488,\n", + " '##UL': 0.08819267818570828,\n", + " '12': 0.08319295152975612,\n", + " '##L': 0.13304534043947008,\n", + " '53': 0.08036439319058314,\n", + " 'Bat': 0.0781627203933898,\n", + " '13': 0.07480367622434814,\n", + " '##ING': 0.05812388913159779,\n", + " 'O': 0.06583622417433492,\n", + " '##éa': 0.05230588344828021,\n", + " '##sh': 0.03433270961563577,\n", + " 'MO': 0.030683540302786836,\n", + " '##imo': 0.017082182198960103,\n", + " 'let': -0.0461111572568852,\n", + " 'remarquable': 0.21263713433629308,\n", + " 'qui': 0.050173365888367016,\n", + " 'bataille': 0.13091439215231107,\n", + " 'Allemagne': 0.11618493617005074,\n", + " 'Fu': 0.10846701242333791,\n", + " 'entre': 0.06369520587259789,\n", + " '##EN': 0.09069418279503459,\n", + " 'fut': 0.08918546414477314,\n", + " '##ER': 0.08840867622071771,\n", + " 'mar': 0.08131159281878217,\n", + " '##SH': 0.07880885570732322,\n", + " 'Brand': 0.07534765223550378,\n", + " 's': 0.07214195178952942,\n", + " 'victoire': 0.07404695440084724,\n", + " 'Albert': 0.06782766064004295,\n", + " '##êché': 0.06760324837539838,\n", + " '##se': 0.06694410055617671,\n", + " '##we': 0.06692906138119581,\n", + " '##illet': 0.06506627545649421,\n", + " '##grave': 0.06490679816106117,\n", + " '##eb': 0.061676421437111995,\n", + " 'Saxe': 0.056318485823556044,\n", + " 'blessure': 0.054096736108623616,\n", + " '##fait': 0.05343946030378872,\n", + " 'Hildesheim': 0.05099271687488835,\n", + " '##fin': 0.04989368895071508,\n", + " 'Maurice': 0.04751196968720714,\n", + " '##ourg': 0.04583679368750848,\n", + " '##lante': 0.03983621993726603,\n", + " '7': 0.03584389050476482,\n", + " 'aux': 0.035433950894822305,\n", + " '1553': 0.034349353274837756,\n", + " 'dé': 0.03434818892830247,\n", + " 'mourut': 0.03360980244865721,\n", + " 'év': 0.029168399068651336,\n", + " '##bourg': 0.02789169534010975,\n", + " 'plusieurs': 0.02669902130425908,\n", + " 'donna': 0.025856769861851345,\n", + " '##ade': 0.023830238910252367,\n", + " 'SW': 0.02082288328341326,\n", + " '##ect': 0.020450927253329462,\n", + " 'Lune': 0.02024761291388681,\n", + " 'él': 0.01741279676584186,\n", + " 'jours': 0.016570340432660837,\n", + " 'sang': 0.015613060865029399,\n", + " 'peu': 0.03996102752601043,\n", + " '##eta': 0.01069390934382476,\n", + " 'ach': 0.010388292909415044,\n", + " 'Ju': 0.0073498549683621225,\n", + " '##eur': 0.0019017102256399137,\n", + " 'y': -0.044396462419225395,\n", + " 'Ty': 0.24148741758807102,\n", + " 'eaux': 0.21684370603760367,\n", + " 'situation': 0.10329824944457552,\n", + " 'cité': 0.09052769883978086,\n", + " 'Or': 0.019460845999311357,\n", + " '##esti': 0.03883953302868402,\n", + " 'Port': -0.1040613277987558,\n", + " 'm': 0.05587762491014585,\n", + " 'Cal': 0.07697134262504549,\n", + " 'appelle': 0.07462975062628824,\n", + " 'car': 0.07461062484523039,\n", + " 'port': 0.06465359698878792,\n", + " 'dire': 0.07059837227461703,\n", + " 'cette': 0.08802705291918328,\n", + " '##rr': 0.06022085803296414,\n", + " 'sept': 0.057519390301443825,\n", + " '##nien': 0.05355078021839007,\n", + " 'fondée': 0.0509239663456544,\n", + " 'pur': 0.05064960239352996,\n", + " 'assez': 0.05042410659440811,\n", + " 'Bari': 0.05030515494264842,\n", + " '##vog': 0.04906152561188761,\n", + " 'pré': 0.04549836271489032,\n", + " '##éta': 0.04507297230896154,\n", + " 'aujourd': 0.04447144001423532,\n", + " '##ne': 0.0347530163976298,\n", + " '##hé': 0.03961438526125526,\n", + " 'ensemble': 0.037956283252757204,\n", + " 'hui': 0.03734807021915644,\n", + " 'leurs': 0.03687680836802711,\n", + " 'être': 0.0343762027708024,\n", + " '##ab': 0.032563279819096465,\n", + " 'pou': 0.03124793207062872,\n", + " '##oi': 0.02989229975500687,\n", + " 'c': -0.006530315247605613,\n", + " '##uri': 0.026965622143799262,\n", + " 'Ra': 0.026882539673090518,\n", + " '##voi': 0.01431455878734083,\n", + " 'suivant': 0.02558394558733645,\n", + " '##o': 0.023331497432489353,\n", + " '##i': 0.020948732860884275,\n", + " 'soit': 0.020530716687414827,\n", + " '##tend': 0.020495566508082305,\n", + " 'loin': 0.019550798182729083,\n", + " '##rie': 0.017169840154727146,\n", + " '##ject': 0.017138782093343834,\n", + " 'ora': 0.017038607009944622,\n", + " '##ît': 0.014602155227601584,\n", + " '##êl': 0.014337833594615367,\n", + " '##oso': 0.010042907154746029,\n", + " '##cle': 0.008070864174746364,\n", + " '##este': -0.004142491388013976,\n", + " 'Quo': -0.011600711502630034,\n", + " '##ifier': -0.024313910709926605,\n", + " 'seul': -0.0305767799476541,\n", + " 'accord': -0.0385442098622452,\n", + " 'Amazon': 0.268750975733531,\n", + " '##Y': 0.19620519591858654,\n", + " 'elle': 0.16664352411683878,\n", + " '##UR': 0.1644776678073895,\n", + " '##AT': 0.15286658491534266,\n", + " '##dion': 0.14541624779420775,\n", + " 'Qui': 0.13771435252173025,\n", + " '##éri': 0.06699488014992556,\n", + " '##xos': 0.11031678713869973,\n", + " '##ale': 0.09162023651074921,\n", + " ':': -0.04107125437938059,\n", + " 'grande': 0.2745048084212555,\n", + " '##G': 0.1593574100830241,\n", + " '37': 0.15551673881995645,\n", + " '##NI': 0.13384155651705198,\n", + " '56': 0.11739328894256833,\n", + " '51': 0.10404197401916239,\n", + " 'Turquie': 0.0990410787816931,\n", + " 'Cara': 0.09295436490234411,\n", + " '##manie': 0.08529201860860249,\n", + " 'CO': 0.07632048130911342,\n", + " 'Portugal': 0.2042976668202922,\n", + " 'ruines': 0.20314744837396317,\n", + " 'plaine': 0.16178744164448852,\n", + " 'située': 0.14874643050344483,\n", + " 'sud': 0.11940264716050225,\n", + " 'embouchure': 0.11627573925225723,\n", + " 'fertile': 0.10563735882960068,\n", + " 'ancienne': 0.09576345241389979,\n", + " 'été': 0.09014077147035093,\n", + " '##T': 0.07841500408788923,\n", + " '##ob': 0.07632579567509333,\n", + " '##AL': 0.07609846437262502,\n", + " 'longueur': 0.061323686830543296,\n", + " 'laquelle': 0.06110922821761415,\n", + " '##UB': 0.05967416217649997,\n", + " 'b': 0.057586434212371265,\n", + " 'mura': 0.054736177441230875,\n", + " 'Cet': 0.05395581578984272,\n", + " '##riga': 0.051846567210072046,\n", + " 'Za': 0.04995526964949323,\n", + " 'plus': 0.04112507132732335,\n", + " '##dao': 0.04753728678912817,\n", + " 'S': -0.04602159510840381,\n", + " '##uba': 0.04550942840492348,\n", + " 'temple': 0.044018212177904886,\n", + " 'pr': 0.04163503255558388,\n", + " 'fruits': 0.04113575313188026,\n", + " '##ét': 0.038606207695351155,\n", + " '##ât': 0.03806250021892484,\n", + " 'Tage': 0.03765080030853683,\n", + " '##l': 0.036545777864497975,\n", + " 'ferme': 0.03580818214260307,\n", + " 'soi': 0.040668276790300634,\n", + " 'forti': 0.03549622233435284,\n", + " 'Lisbon': 0.034290688839724325,\n", + " 'Est': 0.034150407829088125,\n", + " 'grain': 0.03124989952518535,\n", + " '##r': 0.013680800966903325,\n", + " 'vin': 0.030361970705734747,\n", + " '##dou': 0.030107317328042935,\n", + " 'fait': 0.024187752198256366,\n", + " '##tr': 0.023949993283441592,\n", + " 'Elle': 0.022927855199012262,\n", + " '##ucha': 0.009319170948366045,\n", + " 'bout': 0.020885765063577755,\n", + " 'vers': 0.020157214084907837,\n", + " 'eu': 0.01603845134764128,\n", + " 'avant': 0.014469418561726512,\n", + " 'Jupiter': 0.01427103838187756,\n", + " 'Au': 0.013620171062788134,\n", + " '##fier': 0.006096813392549551,\n", + " 'Am': 0.005564187960915995,\n", + " '##ème': 0.004959174323142408,\n", + " '##rama': -0.004348877527747782,\n", + " '##ment': -0.0064917887675220295,\n", + " '##illes': -0.007961146859700573,\n", + " 'ouest': 0.286768593993972,\n", + " 'villages': 0.20582618265616082,\n", + " 'canton': 0.1572237472100742,\n", + " '##UE': 0.13830905321433584,\n", + " '##xi': 0.12866311074641193,\n", + " 'quelques': 0.12300477044276094,\n", + " 'México': 0.11818705671271108,\n", + " '##EP': 0.08843851847857245,\n", + " '##aca': 0.08293512707914791,\n", + " '##Q': 0.07529631420873374,\n", + " 'Pan': 0.06891223037247476,\n", + " '##ho': 0.06749528567182805,\n", + " '##éc': 0.06615958296532198,\n", + " '##fer': 0.06354695452212748,\n", + " '##uco': 0.06345670762758723,\n", + " 'ren': 0.05244102948978422,\n", + " '##me': 0.04554978807764583,\n", + " '##OT': 0.03221971618476101,\n", + " 'XI': -0.011949950511950972,\n", + " 'île': 0.15309594891081557,\n", + " 'Java': 0.16105351665895745,\n", + " '##BA': 0.07463413630204671,\n", + " 'côte': 0.15127253738509425,\n", + " 'TU': 0.05645581765532423,\n", + " 'place': 0.11962756867654223,\n", + " '##N': 0.11801784167947217,\n", + " 'Ban': 0.08482843899835783,\n", + " 'came': 0.08336073380554908,\n", + " 'ceinture': 0.06844534842035621,\n", + " 'près': 0.07520853958224895,\n", + " '##ON': 0.07452683419741084,\n", + " 'haut': 0.04823238398842199,\n", + " '130': 0.04487521666602899,\n", + " 'ils': 0.044544039157585216,\n", + " '##tam': 0.04319673648672328,\n", + " 'nu': 0.03745403392222305,\n", + " 'vont': 0.03470655996970495,\n", + " 'toile': 0.03155735621178085,\n", + " 'belle': 0.03147662648632047,\n", + " '5': 0.0313537019882558,\n", + " 'poi': 0.0294521548866359,\n", + " '##ton': 0.027713782580636243,\n", + " 'Ses': 0.024714865988903954,\n", + " '##gnar': 0.018810819002694525,\n", + " 'leur': 0.012756004109524549,\n", + " 'toute': 0.011917089830431075,\n", + " '##ds': 0.00923735338058978,\n", + " '##d': -0.007455893077793125,\n", + " '##lot': 0.0014408876850427973,\n", + " 'tout': -0.003079531645644194,\n", + " 'traf': -0.005273322147678258,\n", + " 'lati': -0.006057162419608609,\n", + " 'portent': -0.010660914154466471,\n", + " '##ique': -0.014304967428463296,\n", + " 'Sue': 0.18807178158587753,\n", + " 'grand': 0.18657103884394427,\n", + " 'sont': 0.09666358953523904,\n", + " '##de': 0.08031207024907842,\n", + " 'portée': 0.07505629073993414,\n", + " '##ar': 0.05511323593526249,\n", + " '##lie': 0.05204283271035915,\n", + " 'Dal': 0.018382978798721323,\n", + " 'SI': -0.0025126387312130413}" ] }, - "execution_count": 52, + "execution_count": 112, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "cls_explainer.visualize()" + "d['Géographie']" ] }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "markdown", + "execution_count": 117, "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "[('ville', 0.468205314323041),\n", + " ('latin', 0.3960803982086443),\n", + " ('capitale', 0.36838594655129425),\n", + " ('Macédoine', 0.3205630652048948),\n", + " ('province', 0.3138687127885894),\n", + " ('comté', 0.29220668870626426),\n", + " ('lieu', 0.288944483446071),\n", + " ('ouest', 0.286768593993972),\n", + " ('grande', 0.2745048084212555),\n", + " ('Amazon', 0.268750975733531),\n", + " ('mer', 0.26868989143550776),\n", + " ('Orient', 0.26844923609680227),\n", + " ('lac', 0.2627689961654854),\n", + " ('Asie', 0.25951725852047475),\n", + " ('Amérique', 0.2577046079084956),\n", + " ('royaume', 0.2573068212004148),\n", + " ('nord', 0.242739838786995),\n", + " ('fleuve', 0.2423772411435995),\n", + " ('grec', 0.24154434016104503),\n", + " ('Ty', 0.24148741758807102),\n", + " ('vallée', 0.24144223688497182),\n", + " ('Naples', 0.23150278005781894),\n", + " ('G', 0.23135720218825717),\n", + " ('petite', 0.21919648886046295),\n", + " ('septentrional', 0.21736956560941825),\n", + " ('eaux', 0.21684370603760367),\n", + " ('Inde', 0.2133903020411604),\n", + " ('remarquable', 0.21263713433629308),\n", + " ('villages', 0.20582618265616082),\n", + " ('bourg', 0.205192644118189),\n", + " ('##er', 0.20470878331452866),\n", + " ('Portugal', 0.2042976668202922),\n", + " ('ruines', 0.20314744837396317),\n", + " ('distingue', 0.20225033977383755),\n", + " ('##Y', 0.19620519591858654),\n", + " ('nommée', 0.19227297918896974),\n", + " ('Frontera', 0.19152877305938218),\n", + " ('##éo', 0.18973738733606565),\n", + " ('Sue', 0.18807178158587753),\n", + " ('grand', 0.18657103884394427),\n", + " ('rivier', 0.1835301313877702),\n", + " ('##Z', 0.1830676183440645),\n", + " ('milles', 0.18218276737467964),\n", + " ('nom', 0.18136152655341967),\n", + " ('Irlande', 0.17055004990277814),\n", + " ('elle', 0.16664352411683878),\n", + " ('##GE', 0.16634948535987606),\n", + " ('##RI', 0.16460252115303084),\n", + " ('##UR', 0.1644776678073895),\n", + " ('##gr', 0.16202194372565118),\n", + " ('plaine', 0.16178744164448852),\n", + " ('Java', 0.16105351665895745),\n", + " ('##G', 0.1593574100830241),\n", + " ('duché', 0.15878063729951344),\n", + " ('##A', 0.15877744098706795),\n", + " ('partie', 0.15845359313084062),\n", + " ('Italie', 0.1575315512120412),\n", + " ('##J', 0.1573522788542329),\n", + " ('canton', 0.1572237472100742),\n", + " ('37', 0.15551673881995645),\n", + " ('île', 0.15309594891081557),\n", + " ('##AT', 0.15286658491534266),\n", + " ('lieux', 0.15247473369694478),\n", + " ('côte', 0.15127253738509425),\n", + " ('desert', 0.15074697961858496),\n", + " ('située', 0.14874643050344483),\n", + " ('Dublin', 0.14585461747071327),\n", + " ('##dion', 0.14541624779420775),\n", + " ('##I', 0.141653889530405),\n", + " ('##TI', 0.14030883450630402),\n", + " ('##UE', 0.13830905321433584),\n", + " ('Qui', 0.13771435252173025),\n", + " ('habita', 0.13771036963333821),\n", + " ('petit', 0.1376690268440502),\n", + " ('##NI', 0.13384155651705198),\n", + " ('##L', 0.13304534043947008),\n", + " ('bataille', 0.13091439215231107),\n", + " ('Cam', 0.12958328223340024),\n", + " ('##xi', 0.12866311074641193),\n", + " ('##O', 0.12780809591110026),\n", + " ('endroit', 0.12779042183921233),\n", + " ('Novo', 0.12531683029473817),\n", + " ('32', 0.12352770141613803),\n", + " ('##E', 0.12333498427020301),\n", + " ('quelques', 0.12300477044276094),\n", + " ('orientale', 0.12299072447402201),\n", + " ('la', 0.12298598711273313),\n", + " ('empire', 0.12043364298838244),\n", + " ('place', 0.11962756867654223),\n", + " ('sud', 0.11940264716050225),\n", + " ('México', 0.11818705671271108),\n", + " ('##N', 0.11801784167947217),\n", + " ('golf', 0.11792822736721131),\n", + " ('56', 0.11739328894256833),\n", + " ('##ès', 0.11698138135400395),\n", + " ('embouchure', 0.11627573925225723),\n", + " ('Allemagne', 0.11618493617005074),\n", + " ('28', 0.11327727914116813),\n", + " ('##g', 0.11139228310919369),\n", + " ('##xos', 0.11031678713869973),\n", + " ('##MP', 0.10974714271316263),\n", + " ('Fu', 0.10846701242333791),\n", + " ('40', 0.10679936658542148),\n", + " ('fertile', 0.10563735882960068),\n", + " ('depuis', 0.1051878315897289),\n", + " ('Russie', 0.10463858590067449),\n", + " ('51', 0.10404197401916239),\n", + " ('des', 0.10395928916105898),\n", + " ('situation', 0.10329824944457552),\n", + " ('dans', 0.10022254602904804),\n", + " ('Turquie', 0.0990410787816931),\n", + " ('Espagne', 0.09686731766445605),\n", + " ('sont', 0.09666358953523904),\n", + " ('ancienne', 0.09576345241389979),\n", + " ('##c', 0.0948633585054219),\n", + " ('##ce', 0.09317071188401634),\n", + " ('Cara', 0.09295436490234411),\n", + " ('##ale', 0.09162023651074921),\n", + " ('forte', 0.09134790101615797),\n", + " ('##EN', 0.09069418279503459),\n", + " ('cité', 0.09052769883978086),\n", + " ('Gal', 0.09024113726333705),\n", + " ('été', 0.09014077147035093),\n", + " ('commun', 0.08994387122081324),\n", + " ('West', 0.08993142493114488),\n", + " ('ou', 0.08969930275580645),\n", + " ('41', 0.08953747205002806),\n", + " ('##chin', 0.0894472877928192),\n", + " ('fut', 0.08918546414477314),\n", + " ('voir', 0.08847236372959857),\n", + " ('##EP', 0.08843851847857245),\n", + " ('##ER', 0.08840867622071771),\n", + " ('##UL', 0.08819267818570828),\n", + " ('cette', 0.08802705291918328),\n", + " ('se', 0.08786898558747075),\n", + " (',', 0.0858046931906998),\n", + " ('Guadalajara', 0.0857800140375107),\n", + " ('##manie', 0.08529201860860249),\n", + " ('Ban', 0.08482843899835783),\n", + " ('came', 0.08336073380554908),\n", + " ('12', 0.08319295152975612),\n", + " ('##gor', 0.08311355621144252),\n", + " ('à ', 0.08298650083974059),\n", + " ('##aca', 0.08293512707914791),\n", + " ('jet', 0.0829284321696537),\n", + " ('nouvelle', 0.0828520285184145),\n", + " ('##US', 0.0822458506910208),\n", + " ('mar', 0.08131159281878217),\n", + " ('terre', 0.08085795705733743),\n", + " ('##IA', 0.08076525751909835),\n", + " ('53', 0.08036439319058314),\n", + " ('##de', 0.08031207024907842),\n", + " ('deux', 0.07898332743727765),\n", + " ('##SH', 0.07880885570732322),\n", + " ('Allemands', 0.0787047715309721),\n", + " ('##T', 0.07841500408788923),\n", + " ('Bat', 0.0781627203933898),\n", + " ('Co', 0.07783219273122367),\n", + " ('Labour', 0.07756492814166974),\n", + " ('l', 0.07719865579756106),\n", + " ('Cal', 0.07697134262504549),\n", + " ('##ob', 0.07632579567509333),\n", + " ('CO', 0.07632048130911342),\n", + " ('##chine', 0.07614065102466291),\n", + " ('##AL', 0.07609846437262502),\n", + " ('les', 0.0757883643580965),\n", + " ('J', 0.07546336834323314),\n", + " ('Brand', 0.07534765223550378),\n", + " ('##Q', 0.07529631420873374),\n", + " ('près', 0.07520853958224895),\n", + " ('portée', 0.07505629073993414),\n", + " ('13', 0.07480367622434814),\n", + " ('est', 0.074775644174404),\n", + " ('##BA', 0.07463413630204671),\n", + " ('appelle', 0.07462975062628824),\n", + " ('car', 0.07461062484523039),\n", + " ('##ON', 0.07452683419741084),\n", + " ('10', 0.07431163055453546),\n", + " ('victoire', 0.07404695440084724),\n", + " ('##rma', 0.07365816308437048),\n", + " ('romaine', 0.07262679757314543),\n", + " ('s', 0.07214195178952942),\n", + " ('fonte', 0.07176335280650341),\n", + " ('roi', 0.0716659059875927),\n", + " ('##É', 0.07159170704085929),\n", + " ('Da', 0.0714045742494807),\n", + " ('midi', 0.07113877841272949),\n", + " ('##p', 0.07104594769838767),\n", + " ('dire', 0.07059837227461703),\n", + " ('une', 0.0705267038169397),\n", + " ('##AR', 0.0700853626124515),\n", + " ('Pan', 0.06891223037247476),\n", + " ('Long', 0.06868800126003936),\n", + " ('ceinture', 0.06844534842035621),\n", + " ('Albert', 0.06782766064004295),\n", + " ('de', 0.06770303542758496),\n", + " ('##êché', 0.06760324837539838),\n", + " ('##ho', 0.06749528567182805),\n", + " ('neige', 0.06716279133677414),\n", + " ('##éri', 0.06699488014992556),\n", + " ('##se', 0.06694410055617671),\n", + " ('##we', 0.06692906138119581),\n", + " ('##ice', 0.06692411701657223),\n", + " ('##éc', 0.06615958296532198),\n", + " ('O', 0.06583622417433492),\n", + " ('Ce', 0.06555545762824336),\n", + " ('comme', 0.06549376453090695),\n", + " ('##illet', 0.06506627545649421),\n", + " ('##grave', 0.06490679816106117),\n", + " ('A', 0.06479200165315122),\n", + " ('-', 0.06475762097649433),\n", + " ('port', 0.06465359698878792),\n", + " ('##mée', 0.0646449194613384),\n", + " ('entre', 0.06369520587259789),\n", + " ('##fer', 0.06354695452212748),\n", + " ('##uco', 0.06345670762758723),\n", + " ('##eb', 0.061676421437111995),\n", + " ('##ïque', 0.061645981622394784),\n", + " (\"'\", 0.061565814719079584),\n", + " ('longueur', 0.061323686830543296),\n", + " ('laquelle', 0.06110922821761415),\n", + " ('par', 0.061077180857349866),\n", + " ('##bo', 0.06026845472373662),\n", + " ('##rr', 0.06022085803296414),\n", + " ('##UB', 0.05967416217649997),\n", + " ('##vat', 0.05917044284419409),\n", + " ('##ING', 0.05812388913159779),\n", + " ('##te', 0.05775689209126553),\n", + " ('b', 0.057586434212371265),\n", + " ('ori', 0.05756818718647703),\n", + " ('autre', 0.057531150082004046),\n", + " ('ce', 0.057519719498181784),\n", + " ('sept', 0.057519390301443825),\n", + " ('##én', 0.057367966434507346),\n", + " ('toujours', 0.05654656086498567),\n", + " ('TU', 0.05645581765532423),\n", + " ('le', 0.05642199866799782),\n", + " ('Saxe', 0.056318485823556044),\n", + " ('mais', 0.05591848646995751),\n", + " ('m', 0.05587762491014585),\n", + " ('d', 0.05567218618641231),\n", + " ('Ali', 0.055659756416669474),\n", + " ('##ar', 0.05511323593526249),\n", + " ('.', 0.054787505866608105),\n", + " ('mura', 0.054736177441230875),\n", + " ('mod', 0.05471099375321398),\n", + " ('T', 0.0541465558097154),\n", + " ('blessure', 0.054096736108623616),\n", + " ('nomme', 0.05401793892863615),\n", + " ('Cet', 0.05395581578984272),\n", + " ('##nien', 0.05355078021839007),\n", + " ('##fait', 0.05343946030378872),\n", + " ('Lo', 0.05322826492644412),\n", + " ('lat', 0.05319105448357667),\n", + " ('ren', 0.05244102948978422),\n", + " ('##éa', 0.05230588344828021),\n", + " ('30', 0.05213470124947088),\n", + " ('##lie', 0.05204283271035915),\n", + " ('##riga', 0.051846567210072046),\n", + " ('##cci', 0.05157091008708276),\n", + " ('La', 0.051446794552354624),\n", + " ('Lu', 0.05137836438835051),\n", + " ('ensuite', 0.051200901865576555),\n", + " ('Hildesheim', 0.05099271687488835),\n", + " ('fondée', 0.0509239663456544),\n", + " ('##uat', 0.0508369206389474),\n", + " ('où', 0.05080575319876235),\n", + " ('pur', 0.05064960239352996),\n", + " ('CA', 0.05046508890359498),\n", + " ('assez', 0.05042410659440811),\n", + " ('Bari', 0.05030515494264842),\n", + " ('qui', 0.050173365888367016),\n", + " ('grandes', 0.0500708177588014),\n", + " ('Za', 0.04995526964949323),\n", + " ('##fin', 0.04989368895071508),\n", + " ('ses', 0.049645402358164066),\n", + " ('##ii', 0.049587268182288424),\n", + " ('république', 0.04943046233577988),\n", + " ('sur', 0.04915812925999788),\n", + " ('##vog', 0.04906152561188761),\n", + " ('##olo', 0.04880284359551007),\n", + " ('haut', 0.04823238398842199),\n", + " ('##s', 0.047794961185982546),\n", + " ('##dao', 0.04753728678912817),\n", + " ('Maurice', 0.04751196968720714),\n", + " ('##ourg', 0.04583679368750848),\n", + " ('##me', 0.04554978807764583),\n", + " ('##uba', 0.04550942840492348),\n", + " ('pré', 0.04549836271489032),\n", + " ('##éta', 0.04507297230896154),\n", + " ('130', 0.04487521666602899),\n", + " ('##n', 0.04476191426913992),\n", + " ('ils', 0.044544039157585216),\n", + " ('aujourd', 0.04447144001423532),\n", + " ('fort', 0.04442544031371985),\n", + " ('temple', 0.044018212177904886),\n", + " ('##ge', 0.04400356682753303),\n", + " ('M', 0.04357113158723378),\n", + " ('##que', 0.04353082158018988),\n", + " ('va', 0.04337036149864036),\n", + " ('##a', 0.04319994977377055),\n", + " ('##tam', 0.04319673648672328),\n", + " ('C', 0.04306076808452276),\n", + " ('Pt', 0.04241563924208806),\n", + " ('##rab', 0.0423110927679243),\n", + " ('##rée', 0.042257144545369996),\n", + " ('##iot', 0.04203676192154972),\n", + " ('The', 0.04191806043439661),\n", + " ('aussi', 0.04182030961442956),\n", + " ('moderne', 0.04177711856360317),\n", + " ('pr', 0.04163503255558388),\n", + " ('##rro', 0.041311241708329303),\n", + " ('(', 0.0413040601926874),\n", + " ('St', 0.04123715089841821),\n", + " ('an', 0.041146473695852476),\n", + " ('fruits', 0.04113575313188026),\n", + " ('plus', 0.04112507132732335),\n", + " ('##ché', 0.040774048966399426),\n", + " ('o', 0.04070060977171756),\n", + " ('soi', 0.040668276790300634),\n", + " ('X', 0.04064410422404537),\n", + " ('peu', 0.03996102752601043),\n", + " ('Per', 0.03985223606785319),\n", + " ('##lante', 0.03983621993726603),\n", + " ('##hé', 0.03961438526125526),\n", + " ('##j', 0.03926027199963966),\n", + " ('##ée', 0.039227260949307896),\n", + " ('##esti', 0.03883953302868402),\n", + " ('##ét', 0.038606207695351155),\n", + " ('##jana', 0.03818439831225535),\n", + " ('##ima', 0.03813331812513627),\n", + " ('##ât', 0.03806250021892484),\n", + " ('ensemble', 0.037956283252757204),\n", + " ('STAR', 0.037760347726659545),\n", + " ('Tage', 0.03765080030853683),\n", + " ('nu', 0.03745403392222305),\n", + " ('hui', 0.03734807021915644),\n", + " ('Le', 0.03715890519752807),\n", + " ('##us', 0.03713196635047464),\n", + " ('leurs', 0.03687680836802711),\n", + " ('cont', 0.03657382982964496),\n", + " ('##l', 0.036545777864497975),\n", + " ('Sir', 0.03642712798689307),\n", + " ('L', 0.03595023525894895),\n", + " ('7', 0.03584389050476482),\n", + " ('ferme', 0.03580818214260307),\n", + " ('forti', 0.03549622233435284),\n", + " ('aux', 0.035433950894822305),\n", + " ('au', 0.034992091061273066),\n", + " ('##ne', 0.0347530163976298),\n", + " ('vont', 0.03470655996970495),\n", + " ('être', 0.0343762027708024),\n", + " ('1553', 0.034349353274837756),\n", + " ('dé', 0.03434818892830247),\n", + " ('##sh', 0.03433270961563577),\n", + " ('Lisbon', 0.034290688839724325),\n", + " ('Est', 0.034150407829088125),\n", + " ('##z', 0.03365499782170345),\n", + " ('##e', 0.03364232916886872),\n", + " ('mourut', 0.03360980244865721),\n", + " ('##toi', 0.03342140264875253),\n", + " ('un', 0.03322588462984368),\n", + " ('P', 0.03272049148432894),\n", + " ('##ab', 0.032563279819096465),\n", + " ('Ar', 0.032402384389004744),\n", + " ('##OT', 0.03221971618476101),\n", + " ('avo', 0.032212276900181),\n", + " ('sa', 0.032102001051091615),\n", + " ('##ae', 0.03185063483930072),\n", + " ('Hong', 0.03178504105089397),\n", + " ('on', 0.031637879100526636),\n", + " ('toile', 0.03155735621178085),\n", + " ('belle', 0.03147662648632047),\n", + " ('##bal', 0.03142378925271683),\n", + " ('5', 0.0313537019882558),\n", + " ('grain', 0.03124989952518535),\n", + " ('pou', 0.03124793207062872),\n", + " ('##ac', 0.031224938425030026),\n", + " ('é', 0.031069540958438446),\n", + " ('##ia', 0.031040617643215574),\n", + " ('R', 0.030949942486234276),\n", + " ('Tra', 0.030808153025433653),\n", + " ('différentes', 0.030718941715107615),\n", + " ('MO', 0.030683540302786836),\n", + " ('2°', 0.030522230876155416),\n", + " ('vin', 0.030361970705734747),\n", + " ('##dou', 0.030107317328042935),\n", + " ('plu', 0.03005047616189665),\n", + " ('##oi', 0.02989229975500687),\n", + " ('##thus', 0.029706139074928077),\n", + " ('poi', 0.0294521548866359),\n", + " ('VII', 0.02928677630182494),\n", + " ('du', 0.02917787047349929),\n", + " ('év', 0.029168399068651336),\n", + " ('exemple', 0.02878307063738609),\n", + " ('##la', 0.028711752575946197),\n", + " ('##hab', 0.02854877671553812),\n", + " ('##rig', 0.028440723473813656),\n", + " ('Il', 0.0284042208786913),\n", + " ('bien', 0.0283913554226912),\n", + " ('##lp', 0.028348485933541584),\n", + " ('##bourg', 0.02789169534010975),\n", + " ('##ton', 0.027713782580636243),\n", + " ('##uri', 0.026965622143799262),\n", + " ('Ra', 0.026882539673090518),\n", + " ('plusieurs', 0.02669902130425908),\n", + " ('en', 0.026650124208107327),\n", + " ('##ssent', 0.026560785640090002),\n", + " ('ar', 0.02655491286531302),\n", + " ('##dia', 0.026233515028085855),\n", + " ('terme', 0.02610804943109468),\n", + " ('##SA', 0.02594775540828238),\n", + " ('##on', 0.025900328149956048),\n", + " ('##men', 0.025864880430748294),\n", + " ('donna', 0.025856769861851345),\n", + " ('con', 0.025748442770215568),\n", + " ('suivant', 0.02558394558733645),\n", + " ('##é', 0.02500576025356923),\n", + " ('Ses', 0.024714865988903954),\n", + " ('##cus', 0.024522117453619832),\n", + " ('fait', 0.024187752198256366),\n", + " ('##tr', 0.023949993283441592),\n", + " ('##ade', 0.023830238910252367),\n", + " ('##ziu', 0.023636246682512362),\n", + " ('celui', 0.023612546696296227),\n", + " ('E', 0.02353268737063088),\n", + " ('##cer', 0.023491982765047097),\n", + " ('ne', 0.023481641713517833),\n", + " ('Sam', 0.023408934631155117),\n", + " ('##o', 0.023331497432489353),\n", + " ('côté', 0.023297246836922872),\n", + " ('##ent', 0.023156820087233715),\n", + " ('Pier', 0.022936152887815472),\n", + " ('Elle', 0.022927855199012262),\n", + " ('##xiu', 0.022805612841692573),\n", + " ('##od', 0.022641329488828195),\n", + " ('n', 0.022605076586881327),\n", + " ('##noi', 0.02239494074557525),\n", + " ('puisque', 0.022114351087644566),\n", + " ('##es', 0.022001482046772214),\n", + " ('cela', 0.021728661127459404),\n", + " ('##al', 0.021472918740770543),\n", + " ('qu', 0.021413621005676176),\n", + " ('##ros', 0.021349833073345013),\n", + " ('dont', 0.02121929633019263),\n", + " ('D', 0.021142184892655142),\n", + " ('##i', 0.020948732860884275),\n", + " ('Ist', 0.020901809554305005),\n", + " ('bout', 0.020885765063577755),\n", + " ('SW', 0.02082288328341326),\n", + " ('soit', 0.020530716687414827),\n", + " ('##tend', 0.020495566508082305),\n", + " ('##ect', 0.020450927253329462),\n", + " ('ex', 0.020417873274345934),\n", + " ('CIA', 0.020303354874778645),\n", + " ('Lune', 0.02024761291388681),\n", + " ('vers', 0.020157214084907837),\n", + " ('##x', 0.02002047701965849),\n", + " ('loin', 0.019550798182729083),\n", + " ('Or', 0.019460845999311357),\n", + " ('après', 0.01941062923428665),\n", + " ('##euse', 0.019031264633545753),\n", + " ('##gnar', 0.018810819002694525),\n", + " ('app', 0.018555258608997136),\n", + " ('dit', 0.018481328341026063),\n", + " ('Dal', 0.018382978798721323),\n", + " ('##u', 0.018374918227908328),\n", + " ('##jano', 0.01808523851267019),\n", + " ('##ie', 0.017838742894118582),\n", + " ('##gent', 0.01748654204829846),\n", + " ('él', 0.01741279676584186),\n", + " ('Sar', 0.01724425985017986),\n", + " ('##rie', 0.017169840154727146),\n", + " ('##ject', 0.017138782093343834),\n", + " ('##imo', 0.017082182198960103),\n", + " ('ora', 0.017038607009944622),\n", + " ('##ule', 0.0166859230453766),\n", + " ('jours', 0.016570340432660837),\n", + " ('eu', 0.01603845134764128),\n", + " ('Dion', 0.015754758531027112),\n", + " ('sang', 0.015613060865029399),\n", + " ('##re', 0.015364885611094212),\n", + " ('##zog', 0.014949231310082899),\n", + " ('U', 0.014654267300384672),\n", + " ('##ît', 0.014602155227601584),\n", + " ('avant', 0.014469418561726512),\n", + " ('commence', 0.014379318563949204),\n", + " ('##êl', 0.014337833594615367),\n", + " ('##voi', 0.01431455878734083),\n", + " ('Jupiter', 0.01427103838187756),\n", + " ('souvent', 0.01412861492800689),\n", + " ('danger', 0.013811198066105809),\n", + " ('##r', 0.013680800966903325),\n", + " ('Au', 0.013620171062788134),\n", + " ('signifie', 0.013575249323296856),\n", + " ('##pt', 0.013195773759531383),\n", + " ('##oe', 0.013082684704750751),\n", + " ('leur', 0.012756004109524549),\n", + " ('##bre', 0.01200562189665623),\n", + " ('toute', 0.011917089830431075),\n", + " ('met', 0.011327151448637373),\n", + " ('##ies', 0.01091580579338499),\n", + " ('##eta', 0.01069390934382476),\n", + " ('##ns', 0.010472519252392662),\n", + " ('torre', 0.010422364080421317),\n", + " ('ach', 0.010388292909415044),\n", + " ('##oso', 0.010042907154746029),\n", + " ('que', 0.009635328008189985),\n", + " ('##éra', 0.009599163212087692),\n", + " ('##ucha', 0.009319170948366045),\n", + " ('##ds', 0.00923735338058978),\n", + " ('##qui', 0.009204343050603915),\n", + " ('##mi', 0.008906287279368003),\n", + " ('Cassius', 0.008545442448074718),\n", + " ('On', 0.008215765462606122),\n", + " ('##cle', 0.008070864174746364),\n", + " (';', 0.007984840521573309),\n", + " ('Ju', 0.0073498549683621225),\n", + " ('pas', 0.006987930951319245),\n", + " ('co', 0.006828431290704029),\n", + " ('cha', 0.0067258655884391136),\n", + " ('##zet', 0.0063275438230979185),\n", + " ('fine', 0.006245003686740857),\n", + " ('##fier', 0.006096813392549551),\n", + " ('##uy', 0.006086328817223191),\n", + " ('##el', 0.0060696145131198575),\n", + " ('##éir', 0.0059480290294971795),\n", + " ('che', 0.005705843958191305),\n", + " ('ca', 0.0056714294799304005),\n", + " ('Am', 0.005564187960915995),\n", + " ('##roi', 0.005468227545142956),\n", + " ('##dent', 0.00517282162953671),\n", + " ('##ème', 0.004959174323142408),\n", + " ('##crit', 0.004866795460923155),\n", + " ('##t', 0.0048467472813908796),\n", + " ('III', 0.00283531432530362),\n", + " ('trés', 0.002824228104975018),\n", + " ('il', 0.0027654637201536968),\n", + " ('pour', 0.0021363265380875763),\n", + " ('##mon', 0.002094276853030689),\n", + " ('##eur', 0.0019017102256399137),\n", + " ('##oit', 0.001711933310961672),\n", + " ('##bu', 0.0016690113836018768),\n", + " ('##it', 0.0015963173159964771),\n", + " ('##lot', 0.0014408876850427973),\n", + " ('termine', 0.0013832185393301071),\n", + " ('x', 0.0008492312750708209),\n", + " ('[CLS]', 0.0),\n", + " ('[SEP]', 0.0),\n", + " ('sous', -0.0002069181948599748),\n", + " ('a', -0.0012715805160076403),\n", + " ('##ors', -0.0017502978171956062),\n", + " ('SI', -0.0025126387312130413),\n", + " ('tout', -0.003079531645644194),\n", + " ('in', -0.003095317243436783),\n", + " ('donne', -0.0033814500703101104),\n", + " ('##este', -0.004142491388013976),\n", + " ('##rama', -0.004348877527747782),\n", + " ('traf', -0.005273322147678258),\n", + " ('lati', -0.006057162419608609),\n", + " ('##ment', -0.0064917887675220295),\n", + " ('c', -0.006530315247605613),\n", + " ('##d', -0.007455893077793125),\n", + " ('creu', -0.007475521658293897),\n", + " ('h', -0.007529062495164878),\n", + " ('##vo', -0.007773218105616921),\n", + " ('##illes', -0.007961146859700573),\n", + " ('##nt', -0.010146726047092878),\n", + " ('portent', -0.010660914154466471),\n", + " ('Quo', -0.011600711502630034),\n", + " ('XI', -0.011949950511950972),\n", + " ('##ique', -0.014304967428463296),\n", + " ('liv', -0.01661384524933283),\n", + " ('##ades', -0.01685895398059224),\n", + " ('selon', -0.020082683114963043),\n", + " ('##ifier', -0.024313910709926605),\n", + " ('1°', -0.026381150057445416),\n", + " ('##ure', -0.029633249072781927),\n", + " ('seul', -0.0305767799476541),\n", + " ('tem', -0.03507853370176394),\n", + " ('accord', -0.0385442098622452),\n", + " (':', -0.04107125437938059),\n", + " ('born', -0.041898895862637786),\n", + " ('to', -0.043587086241707595),\n", + " ('y', -0.044396462419225395),\n", + " ('audience', -0.045768989771362824),\n", + " ('S', -0.04602159510840381),\n", + " ('let', -0.0461111572568852),\n", + " ('sentiment', -0.048171146031505886),\n", + " ('nach', -0.059647272844616725),\n", + " (')', -0.06872271549404119),\n", + " ('SAR', -0.07143494474675065),\n", + " ('Torre', -0.07370789174796667),\n", + " ('&', -0.08000801041182604),\n", + " ('Comme', -0.08102775997694327),\n", + " ('Port', -0.1040613277987558)]" + ] + }, + "execution_count": 117, + "metadata": {}, + "output_type": "execute_result" + } + ], "source": [ - "* récupérer les mots positifs par domaine (EDdA et LGE)\n", - "* faire des nuages de mots et comparer les plus fréquents entre EDdA et LGE (corpus parallèle)" + "\n", + "l = [(k, v) for k, v in d['Géographie'].items()]\n", + "l.sort(key=lambda a: a[1], reverse = True)\n", + "l\n" ] }, { - "cell_type": "markdown", + "cell_type": "code", + "execution_count": null, "metadata": {}, + "outputs": [], "source": [] } ],