code/example_model/ner_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "pd.set_option('display.max_columns', 500)\n",
    "from transformers import CamembertTokenizerFast, CamembertForTokenClassification, Trainer, TrainingArguments, AdamW\n",
    "from torch.utils.data import DataLoader\n",
    "import torch\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tqdm import tqdm\n",
    "\n",
    "from model_utils import transform_label, truncate, encode_tags, OurDataset, evaluate_model, predict_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "SENTENCES_SIZE = 128\n",
    "TOKENS_NUMBER = 256"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Export des données augmentées"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Token indices sequence length is longer than the specified maximum sequence length for this model (794 > 512). Running this sequence through the model will result in indexing errors\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>35</th>\n",
       "      <th>36</th>\n",
       "      <th>37</th>\n",
       "      <th>38</th>\n",
       "      <th>39</th>\n",
       "      <th>40</th>\n",
       "      <th>41</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "      <th>51</th>\n",
       "      <th>52</th>\n",
       "      <th>53</th>\n",
       "      <th>54</th>\n",
       "      <th>55</th>\n",
       "      <th>56</th>\n",
       "      <th>57</th>\n",
       "      <th>58</th>\n",
       "      <th>59</th>\n",
       "      <th>60</th>\n",
       "      <th>61</th>\n",
       "      <th>62</th>\n",
       "      <th>63</th>\n",
       "      <th>64</th>\n",
       "      <th>65</th>\n",
       "      <th>66</th>\n",
       "      <th>67</th>\n",
       "      <th>68</th>\n",
       "      <th>69</th>\n",
       "      <th>70</th>\n",
       "      <th>71</th>\n",
       "      <th>72</th>\n",
       "      <th>73</th>\n",
       "      <th>74</th>\n",
       "      <th>75</th>\n",
       "      <th>76</th>\n",
       "      <th>77</th>\n",
       "      <th>78</th>\n",
       "      <th>79</th>\n",
       "      <th>80</th>\n",
       "      <th>81</th>\n",
       "      <th>82</th>\n",
       "      <th>83</th>\n",
       "      <th>84</th>\n",
       "      <th>85</th>\n",
       "      <th>86</th>\n",
       "      <th>87</th>\n",
       "      <th>88</th>\n",
       "      <th>89</th>\n",
       "      <th>90</th>\n",
       "      <th>91</th>\n",
       "      <th>92</th>\n",
       "      <th>93</th>\n",
       "      <th>94</th>\n",
       "      <th>95</th>\n",
       "      <th>96</th>\n",
       "      <th>97</th>\n",
       "      <th>98</th>\n",
       "      <th>99</th>\n",
       "      <th>100</th>\n",
       "      <th>101</th>\n",
       "      <th>102</th>\n",
       "      <th>103</th>\n",
       "      <th>104</th>\n",
       "      <th>105</th>\n",
       "      <th>106</th>\n",
       "      <th>107</th>\n",
       "      <th>108</th>\n",
       "      <th>109</th>\n",
       "      <th>110</th>\n",
       "      <th>111</th>\n",
       "      <th>112</th>\n",
       "      <th>113</th>\n",
       "      <th>114</th>\n",
       "      <th>115</th>\n",
       "      <th>116</th>\n",
       "      <th>117</th>\n",
       "      <th>118</th>\n",
       "      <th>119</th>\n",
       "      <th>120</th>\n",
       "      <th>121</th>\n",
       "      <th>122</th>\n",
       "      <th>123</th>\n",
       "      <th>124</th>\n",
       "      <th>125</th>\n",
       "      <th>126</th>\n",
       "      <th>127</th>\n",
       "      <th>128</th>\n",
       "      <th>129</th>\n",
       "      <th>130</th>\n",
       "      <th>131</th>\n",
       "      <th>132</th>\n",
       "      <th>133</th>\n",
       "      <th>134</th>\n",
       "      <th>135</th>\n",
       "      <th>136</th>\n",
       "      <th>137</th>\n",
       "      <th>138</th>\n",
       "      <th>139</th>\n",
       "      <th>140</th>\n",
       "      <th>141</th>\n",
       "      <th>142</th>\n",
       "      <th>143</th>\n",
       "      <th>144</th>\n",
       "      <th>145</th>\n",
       "      <th>146</th>\n",
       "      <th>147</th>\n",
       "      <th>148</th>\n",
       "      <th>149</th>\n",
       "      <th>150</th>\n",
       "      <th>151</th>\n",
       "      <th>152</th>\n",
       "      <th>153</th>\n",
       "      <th>154</th>\n",
       "      <th>155</th>\n",
       "      <th>156</th>\n",
       "      <th>157</th>\n",
       "      <th>158</th>\n",
       "      <th>159</th>\n",
       "      <th>160</th>\n",
       "      <th>161</th>\n",
       "      <th>162</th>\n",
       "      <th>163</th>\n",
       "      <th>164</th>\n",
       "      <th>165</th>\n",
       "      <th>166</th>\n",
       "      <th>167</th>\n",
       "      <th>168</th>\n",
       "      <th>169</th>\n",
       "      <th>170</th>\n",
       "      <th>171</th>\n",
       "      <th>172</th>\n",
       "      <th>173</th>\n",
       "      <th>174</th>\n",
       "      <th>175</th>\n",
       "      <th>176</th>\n",
       "      <th>177</th>\n",
       "      <th>178</th>\n",
       "      <th>179</th>\n",
       "      <th>180</th>\n",
       "      <th>181</th>\n",
       "      <th>182</th>\n",
       "      <th>183</th>\n",
       "      <th>184</th>\n",
       "      <th>185</th>\n",
       "      <th>186</th>\n",
       "      <th>187</th>\n",
       "      <th>188</th>\n",
       "      <th>189</th>\n",
       "      <th>190</th>\n",
       "      <th>191</th>\n",
       "      <th>192</th>\n",
       "      <th>193</th>\n",
       "      <th>194</th>\n",
       "      <th>195</th>\n",
       "      <th>196</th>\n",
       "      <th>197</th>\n",
       "      <th>198</th>\n",
       "      <th>199</th>\n",
       "      <th>200</th>\n",
       "      <th>201</th>\n",
       "      <th>202</th>\n",
       "      <th>203</th>\n",
       "      <th>204</th>\n",
       "      <th>205</th>\n",
       "      <th>206</th>\n",
       "      <th>207</th>\n",
       "      <th>208</th>\n",
       "      <th>209</th>\n",
       "      <th>210</th>\n",
       "      <th>211</th>\n",
       "      <th>212</th>\n",
       "      <th>213</th>\n",
       "      <th>214</th>\n",
       "      <th>215</th>\n",
       "      <th>216</th>\n",
       "      <th>217</th>\n",
       "      <th>218</th>\n",
       "      <th>219</th>\n",
       "      <th>220</th>\n",
       "      <th>221</th>\n",
       "      <th>222</th>\n",
       "      <th>223</th>\n",
       "      <th>224</th>\n",
       "      <th>225</th>\n",
       "      <th>226</th>\n",
       "      <th>227</th>\n",
       "      <th>228</th>\n",
       "      <th>229</th>\n",
       "      <th>230</th>\n",
       "      <th>231</th>\n",
       "      <th>232</th>\n",
       "      <th>233</th>\n",
       "      <th>234</th>\n",
       "      <th>235</th>\n",
       "      <th>236</th>\n",
       "      <th>237</th>\n",
       "      <th>238</th>\n",
       "      <th>239</th>\n",
       "      <th>240</th>\n",
       "      <th>241</th>\n",
       "      <th>242</th>\n",
       "      <th>243</th>\n",
       "      <th>244</th>\n",
       "      <th>245</th>\n",
       "      <th>246</th>\n",
       "      <th>247</th>\n",
       "      <th>248</th>\n",
       "      <th>249</th>\n",
       "      <th>...</th>\n",
       "      <th>56235</th>\n",
       "      <th>56236</th>\n",
       "      <th>56237</th>\n",
       "      <th>56238</th>\n",
       "      <th>56239</th>\n",
       "      <th>56240</th>\n",
       "      <th>56241</th>\n",
       "      <th>56242</th>\n",
       "      <th>56243</th>\n",
       "      <th>56244</th>\n",
       "      <th>56245</th>\n",
       "      <th>56246</th>\n",
       "      <th>56247</th>\n",
       "      <th>56248</th>\n",
       "      <th>56249</th>\n",
       "      <th>56250</th>\n",
       "      <th>56251</th>\n",
       "      <th>56252</th>\n",
       "      <th>56253</th>\n",
       "      <th>56254</th>\n",
       "      <th>56255</th>\n",
       "      <th>56256</th>\n",
       "      <th>56257</th>\n",
       "      <th>56258</th>\n",
       "      <th>56259</th>\n",
       "      <th>56260</th>\n",
       "      <th>56261</th>\n",
       "      <th>56262</th>\n",
       "      <th>56263</th>\n",
       "      <th>56264</th>\n",
       "      <th>56265</th>\n",
       "      <th>56266</th>\n",
       "      <th>56267</th>\n",
       "      <th>56268</th>\n",
       "      <th>56269</th>\n",
       "      <th>56270</th>\n",
       "      <th>56271</th>\n",
       "      <th>56272</th>\n",
       "      <th>56273</th>\n",
       "      <th>56274</th>\n",
       "      <th>56275</th>\n",
       "      <th>56276</th>\n",
       "      <th>56277</th>\n",
       "      <th>56278</th>\n",
       "      <th>56279</th>\n",
       "      <th>56280</th>\n",
       "      <th>56281</th>\n",
       "      <th>56282</th>\n",
       "      <th>56283</th>\n",
       "      <th>56284</th>\n",
       "      <th>56285</th>\n",
       "      <th>56286</th>\n",
       "      <th>56287</th>\n",
       "      <th>56288</th>\n",
       "      <th>56289</th>\n",
       "      <th>56290</th>\n",
       "      <th>56291</th>\n",
       "      <th>56292</th>\n",
       "      <th>56293</th>\n",
       "      <th>56294</th>\n",
       "      <th>56295</th>\n",
       "      <th>56296</th>\n",
       "      <th>56297</th>\n",
       "      <th>56298</th>\n",
       "      <th>56299</th>\n",
       "      <th>56300</th>\n",
       "      <th>56301</th>\n",
       "      <th>56302</th>\n",
       "      <th>56303</th>\n",
       "      <th>56304</th>\n",
       "      <th>56305</th>\n",
       "      <th>56306</th>\n",
       "      <th>56307</th>\n",
       "      <th>56308</th>\n",
       "      <th>56309</th>\n",
       "      <th>56310</th>\n",
       "      <th>56311</th>\n",
       "      <th>56312</th>\n",
       "      <th>56313</th>\n",
       "      <th>56314</th>\n",
       "      <th>56315</th>\n",
       "      <th>56316</th>\n",
       "      <th>56317</th>\n",
       "      <th>56318</th>\n",
       "      <th>56319</th>\n",
       "      <th>56320</th>\n",
       "      <th>56321</th>\n",
       "      <th>56322</th>\n",
       "      <th>56323</th>\n",
       "      <th>56324</th>\n",
       "      <th>56325</th>\n",
       "      <th>56326</th>\n",
       "      <th>56327</th>\n",
       "      <th>56328</th>\n",
       "      <th>56329</th>\n",
       "      <th>56330</th>\n",
       "      <th>56331</th>\n",
       "      <th>56332</th>\n",
       "      <th>56333</th>\n",
       "      <th>56334</th>\n",
       "      <th>56335</th>\n",
       "      <th>56336</th>\n",
       "      <th>56337</th>\n",
       "      <th>56338</th>\n",
       "      <th>56339</th>\n",
       "      <th>56340</th>\n",
       "      <th>56341</th>\n",
       "      <th>56342</th>\n",
       "      <th>56343</th>\n",
       "      <th>56344</th>\n",
       "      <th>56345</th>\n",
       "      <th>56346</th>\n",
       "      <th>56347</th>\n",
       "      <th>56348</th>\n",
       "      <th>56349</th>\n",
       "      <th>56350</th>\n",
       "      <th>56351</th>\n",
       "      <th>56352</th>\n",
       "      <th>56353</th>\n",
       "      <th>56354</th>\n",
       "      <th>56355</th>\n",
       "      <th>56356</th>\n",
       "      <th>56357</th>\n",
       "      <th>56358</th>\n",
       "      <th>56359</th>\n",
       "      <th>56360</th>\n",
       "      <th>56361</th>\n",
       "      <th>56362</th>\n",
       "      <th>56363</th>\n",
       "      <th>56364</th>\n",
       "      <th>56365</th>\n",
       "      <th>56366</th>\n",
       "      <th>56367</th>\n",
       "      <th>56368</th>\n",
       "      <th>56369</th>\n",
       "      <th>56370</th>\n",
       "      <th>56371</th>\n",
       "      <th>56372</th>\n",
       "      <th>56373</th>\n",
       "      <th>56374</th>\n",
       "      <th>56375</th>\n",
       "      <th>56376</th>\n",
       "      <th>56377</th>\n",
       "      <th>56378</th>\n",
       "      <th>56379</th>\n",
       "      <th>56380</th>\n",
       "      <th>56381</th>\n",
       "      <th>56382</th>\n",
       "      <th>56383</th>\n",
       "      <th>56384</th>\n",
       "      <th>56385</th>\n",
       "      <th>56386</th>\n",
       "      <th>56387</th>\n",
       "      <th>56388</th>\n",
       "      <th>56389</th>\n",
       "      <th>56390</th>\n",
       "      <th>56391</th>\n",
       "      <th>56392</th>\n",
       "      <th>56393</th>\n",
       "      <th>56394</th>\n",
       "      <th>56395</th>\n",
       "      <th>56396</th>\n",
       "      <th>56397</th>\n",
       "      <th>56398</th>\n",
       "      <th>56399</th>\n",
       "      <th>56400</th>\n",
       "      <th>56401</th>\n",
       "      <th>56402</th>\n",
       "      <th>56403</th>\n",
       "      <th>56404</th>\n",
       "      <th>56405</th>\n",
       "      <th>56406</th>\n",
       "      <th>56407</th>\n",
       "      <th>56408</th>\n",
       "      <th>56409</th>\n",
       "      <th>56410</th>\n",
       "      <th>56411</th>\n",
       "      <th>56412</th>\n",
       "      <th>56413</th>\n",
       "      <th>56414</th>\n",
       "      <th>56415</th>\n",
       "      <th>56416</th>\n",
       "      <th>56417</th>\n",
       "      <th>56418</th>\n",
       "      <th>56419</th>\n",
       "      <th>56420</th>\n",
       "      <th>56421</th>\n",
       "      <th>56422</th>\n",
       "      <th>56423</th>\n",
       "      <th>56424</th>\n",
       "      <th>56425</th>\n",
       "      <th>56426</th>\n",
       "      <th>56427</th>\n",
       "      <th>56428</th>\n",
       "      <th>56429</th>\n",
       "      <th>56430</th>\n",
       "      <th>56431</th>\n",
       "      <th>56432</th>\n",
       "      <th>56433</th>\n",
       "      <th>56434</th>\n",
       "      <th>56435</th>\n",
       "      <th>56436</th>\n",
       "      <th>56437</th>\n",
       "      <th>56438</th>\n",
       "      <th>56439</th>\n",
       "      <th>56440</th>\n",
       "      <th>56441</th>\n",
       "      <th>56442</th>\n",
       "      <th>56443</th>\n",
       "      <th>56444</th>\n",
       "      <th>56445</th>\n",
       "      <th>56446</th>\n",
       "      <th>56447</th>\n",
       "      <th>56448</th>\n",
       "      <th>56449</th>\n",
       "      <th>56450</th>\n",
       "      <th>56451</th>\n",
       "      <th>56452</th>\n",
       "      <th>56453</th>\n",
       "      <th>56454</th>\n",
       "      <th>56455</th>\n",
       "      <th>56456</th>\n",
       "      <th>56457</th>\n",
       "      <th>56458</th>\n",
       "      <th>56459</th>\n",
       "      <th>56460</th>\n",
       "      <th>56461</th>\n",
       "      <th>56462</th>\n",
       "      <th>56463</th>\n",
       "      <th>56464</th>\n",
       "      <th>56465</th>\n",
       "      <th>56466</th>\n",
       "      <th>56467</th>\n",
       "      <th>56468</th>\n",
       "      <th>56469</th>\n",
       "      <th>56470</th>\n",
       "      <th>56471</th>\n",
       "      <th>56472</th>\n",
       "      <th>56473</th>\n",
       "      <th>56474</th>\n",
       "      <th>56475</th>\n",
       "      <th>56476</th>\n",
       "      <th>56477</th>\n",
       "      <th>56478</th>\n",
       "      <th>56479</th>\n",
       "      <th>56480</th>\n",
       "      <th>56481</th>\n",
       "      <th>56482</th>\n",
       "      <th>56483</th>\n",
       "      <th>56484</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>token</th>\n",
       "      <td>▁Un</td>\n",
       "      <td>▁homme</td>\n",
       "      <td>▁âgé</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁77</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>▁présentait</td>\n",
       "      <td>▁des</td>\n",
       "      <td>▁signes</td>\n",
       "      <td>▁cliniques</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁pro</td>\n",
       "      <td>stat</td>\n",
       "      <td>isme</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Le</td>\n",
       "      <td>▁TR</td>\n",
       "      <td>▁not</td>\n",
       "      <td>ait</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁masse</td>\n",
       "      <td>▁ré</td>\n",
       "      <td>ni</td>\n",
       "      <td>tente</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Le</td>\n",
       "      <td>▁taux</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁PSA</td>\n",
       "      <td>▁était</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁</td>\n",
       "      <td>9,5</td>\n",
       "      <td>▁n</td>\n",
       "      <td>g</td>\n",
       "      <td>/</td>\n",
       "      <td>ml</td>\n",
       "      <td>.</td>\n",
       "      <td>▁L</td>\n",
       "      <td>'</td>\n",
       "      <td>E</td>\n",
       "      <td>ER</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁l</td>\n",
       "      <td>'</td>\n",
       "      <td>imagerie</td>\n",
       "      <td>▁par</td>\n",
       "      <td>▁résonance</td>\n",
       "      <td>▁magnétique</td>\n",
       "      <td>▁nucléaire</td>\n",
       "      <td>▁(</td>\n",
       "      <td>IRM</td>\n",
       "      <td>)</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁prostate</td>\n",
       "      <td>▁mont</td>\n",
       "      <td>raient</td>\n",
       "      <td>▁dans</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁Z</td>\n",
       "      <td>P</td>\n",
       "      <td>▁droite</td>\n",
       "      <td>,</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁masse</td>\n",
       "      <td>▁</td>\n",
       "      <td>ky</td>\n",
       "      <td>s</td>\n",
       "      <td>tique</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁parois</td>\n",
       "      <td>▁épaisse</td>\n",
       "      <td>s</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁4</td>\n",
       "      <td>cm</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁diamètre</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁contenant</td>\n",
       "      <td>▁des</td>\n",
       "      <td>▁végétation</td>\n",
       "      <td>s</td>\n",
       "      <td>▁en</td>\n",
       "      <td>dol</td>\n",
       "      <td>umi</td>\n",
       "      <td>nale</td>\n",
       "      <td>s</td>\n",
       "      <td>.</td>\n",
       "      <td>▁L</td>\n",
       "      <td>'</td>\n",
       "      <td>analyse</td>\n",
       "      <td>▁</td>\n",
       "      <td>histo</td>\n",
       "      <td>logique</td>\n",
       "      <td>▁des</td>\n",
       "      <td>▁bio</td>\n",
       "      <td>ps</td>\n",
       "      <td>ies</td>\n",
       "      <td>▁pro</td>\n",
       "      <td>statique</td>\n",
       "      <td>s</td>\n",
       "      <td>▁écho</td>\n",
       "      <td>gu</td>\n",
       "      <td>idées</td>\n",
       "      <td>▁montrait</td>\n",
       "      <td>▁un</td>\n",
       "      <td>▁ad</td>\n",
       "      <td>éno</td>\n",
       "      <td>car</td>\n",
       "      <td>ci</td>\n",
       "      <td>nome</td>\n",
       "      <td>▁duc</td>\n",
       "      <td>tal</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁score</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁G</td>\n",
       "      <td>le</td>\n",
       "      <td>as</td>\n",
       "      <td>on</td>\n",
       "      <td>▁7</td>\n",
       "      <td>▁(4</td>\n",
       "      <td>+</td>\n",
       "      <td>3)</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁contingent</td>\n",
       "      <td>▁tra</td>\n",
       "      <td>bé</td>\n",
       "      <td>culaire</td>\n",
       "      <td>▁pré</td>\n",
       "      <td>dom</td>\n",
       "      <td>in</td>\n",
       "      <td>ant</td>\n",
       "      <td>.</td>\n",
       "      <td>▁L</td>\n",
       "      <td>'</td>\n",
       "      <td>analyse</td>\n",
       "      <td>▁immuno</td>\n",
       "      <td>-</td>\n",
       "      <td>histo</td>\n",
       "      <td>logique</td>\n",
       "      <td>▁pour</td>\n",
       "      <td>▁les</td>\n",
       "      <td>▁cyto</td>\n",
       "      <td>k</td>\n",
       "      <td>érat</td>\n",
       "      <td>ines</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁haut</td>\n",
       "      <td>▁poids</td>\n",
       "      <td>▁moléculaire</td>\n",
       "      <td>▁(</td>\n",
       "      <td>anti</td>\n",
       "      <td>▁C</td>\n",
       "      <td>K</td>\n",
       "      <td>▁90</td>\n",
       "      <td>3)</td>\n",
       "      <td>▁était</td>\n",
       "      <td>▁négative</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Le</td>\n",
       "      <td>▁stade</td>\n",
       "      <td>▁clinique</td>\n",
       "      <td>▁était</td>\n",
       "      <td>▁lors</td>\n",
       "      <td>▁du</td>\n",
       "      <td>▁diagnostic</td>\n",
       "      <td>▁:</td>\n",
       "      <td>▁T</td>\n",
       "      <td>3</td>\n",
       "      <td>N</td>\n",
       "      <td>x</td>\n",
       "      <td>M</td>\n",
       "      <td>0</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Un</td>\n",
       "      <td>▁traitement</td>\n",
       "      <td>▁par</td>\n",
       "      <td>▁blocage</td>\n",
       "      <td>▁</td>\n",
       "      <td>andro</td>\n",
       "      <td>génique</td>\n",
       "      <td>▁complet</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁été</td>\n",
       "      <td>▁indiqué</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Le</td>\n",
       "      <td>▁PSA</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁6</td>\n",
       "      <td>▁mois</td>\n",
       "      <td>▁était</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁0,3</td>\n",
       "      <td>6</td>\n",
       "      <td>▁n</td>\n",
       "      <td>g</td>\n",
       "      <td>/</td>\n",
       "      <td>ml</td>\n",
       "      <td>▁puis</td>\n",
       "      <td>▁avait</td>\n",
       "      <td>▁augmenté</td>\n",
       "      <td>▁progressivement</td>\n",
       "      <td>▁jusqu</td>\n",
       "      <td>'</td>\n",
       "      <td>à</td>\n",
       "      <td>▁atteindre</td>\n",
       "      <td>▁3</td>\n",
       "      <td>,</td>\n",
       "      <td>2</td>\n",
       "      <td>▁n</td>\n",
       "      <td>g</td>\n",
       "      <td>/</td>\n",
       "      <td>ml</td>\n",
       "      <td>,</td>\n",
       "      <td>▁3</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>▁après</td>\n",
       "      <td>▁le</td>\n",
       "      <td>▁diagnostic</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Une</td>\n",
       "      <td>▁E</td>\n",
       "      <td>ER</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁contrôle</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁3</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁été</td>\n",
       "      <td>▁réalisée</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁montrait</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁persistance</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁cette</td>\n",
       "      <td>▁masse</td>\n",
       "      <td>▁</td>\n",
       "      <td>ky</td>\n",
       "      <td>s</td>\n",
       "      <td>tique</td>\n",
       "      <td>▁mais</td>\n",
       "      <td>▁avec</td>\n",
       "      <td>▁des</td>\n",
       "      <td>...</td>\n",
       "      <td>▁en</td>\n",
       "      <td>▁place</td>\n",
       "      <td>▁associé</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁en</td>\n",
       "      <td>té</td>\n",
       "      <td>roc</td>\n",
       "      <td>y</td>\n",
       "      <td>st</td>\n",
       "      <td>oplastie</td>\n",
       "      <td>▁d</td>\n",
       "      <td>'</td>\n",
       "      <td>élargissement</td>\n",
       "      <td>▁en</td>\n",
       "      <td>▁raison</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁l</td>\n",
       "      <td>'</td>\n",
       "      <td>hypo</td>\n",
       "      <td>can</td>\n",
       "      <td>ie</td>\n",
       "      <td>▁sévère</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Le</td>\n",
       "      <td>▁patient</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁continué</td>\n",
       "      <td>auto</td>\n",
       "      <td>-</td>\n",
       "      <td>son</td>\n",
       "      <td>dage</td>\n",
       "      <td>s</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Ils</td>\n",
       "      <td>▁sont</td>\n",
       "      <td>▁un</td>\n",
       "      <td>▁patient</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁59</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>▁qui</td>\n",
       "      <td>▁s</td>\n",
       "      <td>'</td>\n",
       "      <td>est</td>\n",
       "      <td>▁présenté</td>\n",
       "      <td>▁aux</td>\n",
       "      <td>▁urgences</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁</td>\n",
       "      <td>hémat</td>\n",
       "      <td>urie</td>\n",
       "      <td>▁macro</td>\n",
       "      <td>scopique</td>\n",
       "      <td>▁spontanée</td>\n",
       "      <td>,</td>\n",
       "      <td>▁indo</td>\n",
       "      <td>lore</td>\n",
       "      <td>▁sans</td>\n",
       "      <td>▁fièvre</td>\n",
       "      <td>▁ni</td>\n",
       "      <td>▁perte</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁poids</td>\n",
       "      <td>.</td>\n",
       "      <td>L</td>\n",
       "      <td>'</td>\n",
       "      <td>examen</td>\n",
       "      <td>▁était</td>\n",
       "      <td>▁normal</td>\n",
       "      <td>▁avec</td>\n",
       "      <td>▁des</td>\n",
       "      <td>▁réservoir</td>\n",
       "      <td>s</td>\n",
       "      <td>▁lombaire</td>\n",
       "      <td>s</td>\n",
       "      <td>▁flexible</td>\n",
       "      <td>s</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁sans</td>\n",
       "      <td>▁douleur</td>\n",
       "      <td>.</td>\n",
       "      <td>▁Un</td>\n",
       "      <td>▁75</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>▁-</td>\n",
       "      <td>old</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁été</td>\n",
       "      <td>▁exploité</td>\n",
       "      <td>▁pour</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁première</td>\n",
       "      <td>▁fois</td>\n",
       "      <td>▁avec</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁tumeur</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁vessie</td>\n",
       "      <td>▁superficielle</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁multi</td>\n",
       "      <td>focal</td>\n",
       "      <td>e</td>\n",
       "      <td>,</td>\n",
       "      <td>▁le</td>\n",
       "      <td>▁stade</td>\n",
       "      <td>▁clinique</td>\n",
       "      <td>▁TA</td>\n",
       "      <td>▁G</td>\n",
       "      <td>II</td>\n",
       "      <td>,</td>\n",
       "      <td>▁découvert</td>\n",
       "      <td>▁pendant</td>\n",
       "      <td>▁le</td>\n",
       "      <td>▁système</td>\n",
       "      <td>▁urinaire</td>\n",
       "      <td>▁irr</td>\n",
       "      <td>in</td>\n",
       "      <td>aire</td>\n",
       "      <td>.</td>\n",
       "      <td>contr</td>\n",
       "      <td>ô</td>\n",
       "      <td>lé</td>\n",
       "      <td>▁par</td>\n",
       "      <td>▁un</td>\n",
       "      <td>▁dérivé</td>\n",
       "      <td>▁ni</td>\n",
       "      <td>tro</td>\n",
       "      <td>,</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁hé</td>\n",
       "      <td>mor</td>\n",
       "      <td>r</td>\n",
       "      <td>oïd</td>\n",
       "      <td>ectomie</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁greffe</td>\n",
       "      <td>▁cutanée</td>\n",
       "      <td>▁pour</td>\n",
       "      <td>▁les</td>\n",
       "      <td>▁brûlure</td>\n",
       "      <td>s</td>\n",
       "      <td>▁électriques</td>\n",
       "      <td>▁con</td>\n",
       "      <td>stitu</td>\n",
       "      <td>aient</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁majeure</td>\n",
       "      <td>▁partie</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁son</td>\n",
       "      <td>▁histoire</td>\n",
       "      <td>▁personnelle</td>\n",
       "      <td>.</td>\n",
       "      <td>▁D</td>\n",
       "      <td>'</td>\n",
       "      <td>autre</td>\n",
       "      <td>▁part</td>\n",
       "      <td>,</td>\n",
       "      <td>▁sa</td>\n",
       "      <td>▁mère</td>\n",
       "      <td>▁est</td>\n",
       "      <td>▁décédée</td>\n",
       "      <td>▁d</td>\n",
       "      <td>'</td>\n",
       "      <td>une</td>\n",
       "      <td>▁tuberculose</td>\n",
       "      <td>▁pulmonaire</td>\n",
       "      <td>▁plusieurs</td>\n",
       "      <td>▁décennies</td>\n",
       "      <td>▁auparavant</td>\n",
       "      <td>.</td>\n",
       "      <td>▁A</td>\n",
       "      <td>.</td>\n",
       "      <td>M</td>\n",
       "      <td>.</td>\n",
       "      <td>,</td>\n",
       "      <td>▁53</td>\n",
       "      <td>▁ans</td>\n",
       "      <td>,</td>\n",
       "      <td>▁un</td>\n",
       "      <td>▁grand</td>\n",
       "      <td>▁</td>\n",
       "      <td>fumeur</td>\n",
       "      <td>,</td>\n",
       "      <td>▁sans</td>\n",
       "      <td>▁aucune</td>\n",
       "      <td>▁histoire</td>\n",
       "      <td>▁particulière</td>\n",
       "      <td>,</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁été</td>\n",
       "      <td>▁admis</td>\n",
       "      <td>▁à</td>\n",
       "      <td>▁notre</td>\n",
       "      <td>▁service</td>\n",
       "      <td>▁en</td>\n",
       "      <td>▁juin</td>\n",
       "      <td>▁1998</td>\n",
       "      <td>▁pour</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁</td>\n",
       "      <td>hémat</td>\n",
       "      <td>urie</td>\n",
       "      <td>▁macro</td>\n",
       "      <td>scopique</td>\n",
       "      <td>▁dont</td>\n",
       "      <td>▁l</td>\n",
       "      <td>'</td>\n",
       "      <td>évaluation</td>\n",
       "      <td>▁en</td>\n",
       "      <td>do</td>\n",
       "      <td>scopique</td>\n",
       "      <td>▁a</td>\n",
       "      <td>▁permis</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁découvrir</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁tumeur</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁vessie</td>\n",
       "      <td>▁compacte</td>\n",
       "      <td>▁du</td>\n",
       "      <td>▁d</td>\n",
       "      <td>ôme</td>\n",
       "      <td>▁et</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁la</td>\n",
       "      <td>▁face</td>\n",
       "      <td>▁postérieure</td>\n",
       "      <td>,</td>\n",
       "      <td>▁une</td>\n",
       "      <td>▁grande</td>\n",
       "      <td>▁importance</td>\n",
       "      <td>▁de</td>\n",
       "      <td>▁base</td>\n",
       "      <td>et</td>\n",
       "      <td>▁dure</td>\n",
       "      <td>▁en</td>\n",
       "      <td>▁contact</td>\n",
       "      <td>▁avec</td>\n",
       "      <td>▁le</td>\n",
       "      <td>▁c</td>\n",
       "      <td>y</td>\n",
       "      <td>st</td>\n",
       "      <td>oscope</td>\n",
       "      <td>.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>label</th>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>...</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>Treatment</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "      <td>O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id_phrase</th>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>...</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>559101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>560101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>561101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "      <td>562101000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 954134 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          0       1     2     3     4     5            6     7        8      \\\n",
       "token       ▁Un  ▁homme  ▁âgé   ▁de   ▁77  ▁ans  ▁présentait  ▁des  ▁signes   \n",
       "label         O       O     O     O     O     O            O     O        O   \n",
       "id_phrase  1000    1000  1000  1000  1000  1000         1000  1000     1000   \n",
       "\n",
       "                9     10    11    12    13    14    15    16    17    18     \\\n",
       "token      ▁cliniques   ▁de  ▁pro  stat  isme     .   ▁Le   ▁TR  ▁not   ait   \n",
       "label               O     O     O     O     O     O     O     O     O     O   \n",
       "id_phrase        1000  1000  1000  1000  1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "          19      20    21    22     23    24    25     26    27    28     \\\n",
       "token      ▁une  ▁masse   ▁ré    ni  tente     .   ▁Le  ▁taux   ▁de  ▁PSA   \n",
       "label         O       O     O     O      O     O     O      O     O     O   \n",
       "id_phrase  1000    1000  1000  1000   1000  1000  1000   1000  1000  1000   \n",
       "\n",
       "            29    30    31    32    33    34    35    36    37    38    39     \\\n",
       "token      ▁était   ▁de     ▁   9,5    ▁n     g     /    ml     .    ▁L     '   \n",
       "label           O     O     O     O     O     O     O     O     O     O     O   \n",
       "id_phrase    1000  1000  1000  1000  1000  1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "          40    41    42    43    44        45    46          47     \\\n",
       "token         E    ER   ▁et    ▁l     '  imagerie  ▁par  ▁résonance   \n",
       "label         O     O     O     O     O         O     O           O   \n",
       "id_phrase  1000  1000  1000  1000  1000      1000  1000        1000   \n",
       "\n",
       "                 48          49    50    51    52    53    54         55     \\\n",
       "token      ▁magnétique  ▁nucléaire    ▁(   IRM     )   ▁de   ▁la  ▁prostate   \n",
       "label                O           O     O     O     O     O     O          O   \n",
       "id_phrase         1000        1000  1000  1000  1000  1000  1000       1000   \n",
       "\n",
       "           56      57     58    59    60    61       62    63    64     \\\n",
       "token      ▁mont  raient  ▁dans   ▁la    ▁Z     P  ▁droite     ,  ▁une   \n",
       "label          O       O      O     O     O     O        O     O     O   \n",
       "id_phrase   1000    1000   1000  1000  1000  1000     1000  1000  1000   \n",
       "\n",
       "            65    66    67    68     69    70       71        72    73     \\\n",
       "token      ▁masse     ▁    ky     s  tique    ▁à  ▁parois  ▁épaisse     s   \n",
       "label           O     O     O     O      O     O        O         O     O   \n",
       "id_phrase    1000  1000  1000  1000   1000  1000     1000      1000  1000   \n",
       "\n",
       "          74    75    76    77         78    79          80    81     \\\n",
       "token       ▁de    ▁4    cm   ▁de  ▁diamètre   ▁et  ▁contenant  ▁des   \n",
       "label         O     O     O     O          O     O           O     O   \n",
       "id_phrase  1000  1000  1000  1000       1000  1000        1000  1000   \n",
       "\n",
       "                 82    83    84    85    86    87    88    89    90    91     \\\n",
       "token      ▁végétation     s   ▁en   dol   umi  nale     s     .    ▁L     '   \n",
       "label                O     O     O     O     O     O     O     O     O     O   \n",
       "id_phrase         1000  1000  1000  1000  1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "             92    93     94       95    96    97    98    99    100    \\\n",
       "token      analyse     ▁  histo  logique  ▁des  ▁bio    ps   ies  ▁pro   \n",
       "label            O     O      O        O     O     O     O     O     O   \n",
       "id_phrase     1000  1000   1000     1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "              101   102    103   104    105        106   107   108   109    \\\n",
       "token      statique     s  ▁écho    gu  idées  ▁montrait   ▁un   ▁ad   éno   \n",
       "label             O     O      O     O      O          O     O     O     O   \n",
       "id_phrase      1000  1000   1000  1000   1000       1000  1000  1000  1000   \n",
       "\n",
       "          110   111   112   113   114   115     116   117   118   119   120    \\\n",
       "token       car    ci  nome  ▁duc   tal   ▁de  ▁score   ▁de    ▁G    le    as   \n",
       "label         O     O     O     O     O     O       O     O     O     O     O   \n",
       "id_phrase  1000  1000  1000  1000  1000  1000    1000  1000  1000  1000  1000   \n",
       "\n",
       "          121   122   123   124   125   126   127          128   129   130    \\\n",
       "token        on    ▁7   ▁(4     +    3)   ▁et   ▁de  ▁contingent  ▁tra    bé   \n",
       "label         O     O     O     O     O     O     O            O     O     O   \n",
       "id_phrase  1000  1000  1000  1000  1000  1000  1000         1000  1000  1000   \n",
       "\n",
       "             131   132   133   134   135   136   137   138      139    \\\n",
       "token      culaire  ▁pré   dom    in   ant     .    ▁L     '  analyse   \n",
       "label            O     O     O     O     O     O     O     O        O   \n",
       "id_phrase     1000  1000  1000  1000  1000  1000  1000  1000     1000   \n",
       "\n",
       "             140   141    142      143    144   145    146   147   148    \\\n",
       "token      ▁immuno     -  histo  logique  ▁pour  ▁les  ▁cyto     k  érat   \n",
       "label            O     O      O        O      O     O      O     O     O   \n",
       "id_phrase     1000  1000   1000     1000   1000  1000   1000  1000  1000   \n",
       "\n",
       "          149   150    151     152           153   154   155   156   157    \\\n",
       "token      ines   ▁de  ▁haut  ▁poids  ▁moléculaire    ▁(  anti    ▁C     K   \n",
       "label         O     O      O       O             O     O     O     O     O   \n",
       "id_phrase  1000  1000   1000    1000          1000  1000  1000  1000  1000   \n",
       "\n",
       "          158   159     160        161   162   163     164        165    \\\n",
       "token       ▁90    3)  ▁était  ▁négative     .   ▁Le  ▁stade  ▁clinique   \n",
       "label         O     O       O          O     O     O       O          O   \n",
       "id_phrase  1000  1000    1000       1000  1000  1000    1000       1000   \n",
       "\n",
       "            166    167   168          169   170   171   172   173   174    \\\n",
       "token      ▁était  ▁lors   ▁du  ▁diagnostic    ▁:    ▁T     3     N     x   \n",
       "label           O      O     O            O     O     O     O     O     O   \n",
       "id_phrase    1000   1000  1000         1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "          175   176   177   178          179   180        181        182    \\\n",
       "token         M     0     .   ▁Un  ▁traitement  ▁par   ▁blocage          ▁   \n",
       "label         O     O     O     O            O     O  Treatment  Treatment   \n",
       "id_phrase  1000  1000  1000  1000         1000  1000       1000       1000   \n",
       "\n",
       "               183        184       185   186   187       188   189   190    \\\n",
       "token          andro    génique  ▁complet    ▁a  ▁été  ▁indiqué     .   ▁Le   \n",
       "label      Treatment  Treatment         O     O     O         O     O     O   \n",
       "id_phrase       1000       1000      1000  1000  1000      1000  1000  1000   \n",
       "\n",
       "          191   192   193    194     195   196   197   198   199   200    \\\n",
       "token      ▁PSA    ▁à    ▁6  ▁mois  ▁était    ▁à  ▁0,3     6    ▁n     g   \n",
       "label         O     O     O      O       O     O     O     O     O     O   \n",
       "id_phrase  1000  1000  1000   1000    1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "          201   202    203     204        205               206     207    \\\n",
       "token         /    ml  ▁puis  ▁avait  ▁augmenté  ▁progressivement  ▁jusqu   \n",
       "label         O     O      O       O          O                 O       O   \n",
       "id_phrase  1000  1000   1000    1000       1000              1000    1000   \n",
       "\n",
       "          208   209         210   211   212   213   214   215   216   217    \\\n",
       "token         '     à  ▁atteindre    ▁3     ,     2    ▁n     g     /    ml   \n",
       "label         O     O           O     O     O     O     O     O     O     O   \n",
       "id_phrase  1000  1000        1000  1000  1000  1000  1000  1000  1000  1000   \n",
       "\n",
       "          218   219   220     221   222          223   224   225   226    \\\n",
       "token         ,    ▁3  ▁ans  ▁après   ▁le  ▁diagnostic     .  ▁Une    ▁E   \n",
       "label         O     O     O       O     O            O     O     O     O   \n",
       "id_phrase  1000  1000  1000    1000  1000         1000  1000  1000  1000   \n",
       "\n",
       "          227   228        229   230   231   232   233   234        235    \\\n",
       "token        ER   ▁de  ▁contrôle    ▁à    ▁3  ▁ans    ▁a  ▁été  ▁réalisée   \n",
       "label         O     O          O     O     O     O     O     O          O   \n",
       "id_phrase  1000  1000       1000  1000  1000  1000  1000  1000       1000   \n",
       "\n",
       "          236        237   238           239   240     241     242   243    \\\n",
       "token       ▁et  ▁montrait   ▁la  ▁persistance   ▁de  ▁cette  ▁masse     ▁   \n",
       "label         O          O     O             O     O       O       O     O   \n",
       "id_phrase  1000       1000  1000          1000  1000    1000    1000  1000   \n",
       "\n",
       "          244   245    246    247    248   249    ...      56235      56236  \\\n",
       "token        ky     s  tique  ▁mais  ▁avec  ▁des  ...        ▁en     ▁place   \n",
       "label         O     O      O      O      O     O  ...          O          O   \n",
       "id_phrase  1000  1000   1000   1000   1000  1000  ...  559101000  559101000   \n",
       "\n",
       "               56237      56238      56239      56240      56241      56242  \\\n",
       "token       ▁associé         ▁à       ▁une        ▁en         té        roc   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  559101000  559101000  559101000  559101000  559101000  559101000   \n",
       "\n",
       "               56243      56244      56245      56246      56247  \\\n",
       "token              y         st   oplastie         ▁d          '   \n",
       "label              O          O          O          O          O   \n",
       "id_phrase  559101000  559101000  559101000  559101000  559101000   \n",
       "\n",
       "                   56248      56249      56250      56251      56252  \\\n",
       "token      élargissement        ▁en    ▁raison        ▁de         ▁l   \n",
       "label                  O          O          O          O          O   \n",
       "id_phrase      559101000  559101000  559101000  559101000  559101000   \n",
       "\n",
       "               56253      56254      56255      56256      56257      56258  \\\n",
       "token              '       hypo        can         ie    ▁sévère          .   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  559101000  559101000  559101000  559101000  559101000  559101000   \n",
       "\n",
       "               56259      56260      56261      56262      56263      56264  \\\n",
       "token            ▁Le   ▁patient         ▁a  ▁continué       auto          -   \n",
       "label              O          O          O  Treatment  Treatment  Treatment   \n",
       "id_phrase  559101000  559101000  559101000  559101000  559101000  559101000   \n",
       "\n",
       "               56265      56266      56267      56268      56269      56270  \\\n",
       "token            son       dage          s          .       ▁Ils      ▁sont   \n",
       "label      Treatment  Treatment  Treatment  Treatment          O          O   \n",
       "id_phrase  559101000  559101000  559101000  559101000  560101000  560101000   \n",
       "\n",
       "               56271      56272      56273      56274      56275      56276  \\\n",
       "token            ▁un   ▁patient        ▁de        ▁59       ▁ans       ▁qui   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56277      56278      56279      56280      56281      56282  \\\n",
       "token             ▁s          '        est  ▁présenté       ▁aux  ▁urgences   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56283      56284      56285      56286      56287      56288  \\\n",
       "token             ▁à       ▁une          ▁      hémat       urie     ▁macro   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56289       56290      56291      56292      56293      56294  \\\n",
       "token       scopique  ▁spontanée          ,      ▁indo       lore      ▁sans   \n",
       "label              O           O          O          O          O          O   \n",
       "id_phrase  560101000   560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56295      56296      56297      56298      56299      56300  \\\n",
       "token        ▁fièvre        ▁ni     ▁perte        ▁de     ▁poids          .   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56301      56302      56303      56304      56305      56306  \\\n",
       "token              L          '     examen     ▁était    ▁normal      ▁avec   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56307       56308      56309      56310      56311      56312  \\\n",
       "token           ▁des  ▁réservoir          s  ▁lombaire          s  ▁flexible   \n",
       "label              O           O          O          O          O          O   \n",
       "id_phrase  560101000   560101000  560101000  560101000  560101000  560101000   \n",
       "\n",
       "               56313      56314      56315      56316      56317      56318  \\\n",
       "token              s        ▁et      ▁sans   ▁douleur          .        ▁Un   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  560101000  560101000  560101000  560101000  560101000  561101000   \n",
       "\n",
       "               56319      56320      56321      56322      56323      56324  \\\n",
       "token            ▁75       ▁ans         ▁-        old         ▁a       ▁été   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56325      56326      56327      56328      56329      56330  \\\n",
       "token      ▁exploité      ▁pour        ▁la  ▁première      ▁fois      ▁avec   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56331      56332      56333      56334           56335  \\\n",
       "token           ▁une    ▁tumeur        ▁de    ▁vessie  ▁superficielle   \n",
       "label              O          O          O          O               O   \n",
       "id_phrase  561101000  561101000  561101000  561101000       561101000   \n",
       "\n",
       "               56336      56337      56338      56339      56340      56341  \\\n",
       "token            ▁et     ▁multi      focal          e          ,        ▁le   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56342      56343      56344      56345      56346      56347  \\\n",
       "token         ▁stade  ▁clinique        ▁TA         ▁G         II          ,   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "                56348      56349      56350      56351      56352      56353  \\\n",
       "token      ▁découvert   ▁pendant        ▁le   ▁système  ▁urinaire       ▁irr   \n",
       "label               O          O          O          O          O          O   \n",
       "id_phrase   561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56354      56355      56356      56357      56358      56359  \\\n",
       "token             in       aire          .      contr          ô         lé   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56360      56361      56362      56363      56364      56365  \\\n",
       "token           ▁par        ▁un    ▁dérivé        ▁ni        tro          ,   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56366      56367      56368      56369      56370      56371  \\\n",
       "token           ▁une        ▁hé        mor          r        oïd    ectomie   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56372      56373      56374      56375      56376      56377  \\\n",
       "token            ▁et       ▁une    ▁greffe   ▁cutanée      ▁pour       ▁les   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56378      56379         56380      56381      56382  \\\n",
       "token       ▁brûlure          s  ▁électriques       ▁con      stitu   \n",
       "label              O          O             O          O          O   \n",
       "id_phrase  561101000  561101000     561101000  561101000  561101000   \n",
       "\n",
       "               56383      56384      56385      56386      56387      56388  \\\n",
       "token          aient        ▁la   ▁majeure    ▁partie        ▁de       ▁son   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56389         56390      56391      56392      56393  \\\n",
       "token      ▁histoire  ▁personnelle          .         ▁D          '   \n",
       "label              O             O          O          O          O   \n",
       "id_phrase  561101000     561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56394      56395      56396      56397      56398      56399  \\\n",
       "token          autre      ▁part          ,        ▁sa      ▁mère       ▁est   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  561101000  561101000  561101000  561101000  561101000  561101000   \n",
       "\n",
       "               56400      56401      56402      56403         56404  \\\n",
       "token       ▁décédée         ▁d          '        une  ▁tuberculose   \n",
       "label              O          O          O          O             O   \n",
       "id_phrase  561101000  561101000  561101000  561101000     561101000   \n",
       "\n",
       "                 56405       56406       56407        56408      56409  \\\n",
       "token      ▁pulmonaire  ▁plusieurs  ▁décennies  ▁auparavant          .   \n",
       "label                O           O           O            O          O   \n",
       "id_phrase    561101000   561101000   561101000    561101000  561101000   \n",
       "\n",
       "               56410      56411      56412      56413      56414      56415  \\\n",
       "token             ▁A          .          M          .          ,        ▁53   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56416      56417      56418      56419      56420      56421  \\\n",
       "token           ▁ans          ,        ▁un     ▁grand          ▁     fumeur   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56422      56423      56424      56425          56426  \\\n",
       "token              ,      ▁sans    ▁aucune  ▁histoire  ▁particulière   \n",
       "label              O          O          O          O              O   \n",
       "id_phrase  562101000  562101000  562101000  562101000      562101000   \n",
       "\n",
       "               56427      56428      56429      56430      56431      56432  \\\n",
       "token              ,         ▁a       ▁été     ▁admis         ▁à     ▁notre   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56433      56434      56435      56436      56437      56438  \\\n",
       "token       ▁service        ▁en      ▁juin      ▁1998      ▁pour       ▁une   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56439      56440      56441      56442      56443      56444  \\\n",
       "token              ▁      hémat       urie     ▁macro   scopique      ▁dont   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56445      56446       56447      56448      56449      56450  \\\n",
       "token             ▁l          '  évaluation        ▁en         do   scopique   \n",
       "label              O          O           O          O          O          O   \n",
       "id_phrase  562101000  562101000   562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56451      56452      56453       56454      56455      56456  \\\n",
       "token             ▁a    ▁permis        ▁de  ▁découvrir       ▁une    ▁tumeur   \n",
       "label              O          O          O           O          O          O   \n",
       "id_phrase  562101000  562101000  562101000   562101000  562101000  562101000   \n",
       "\n",
       "               56457      56458      56459      56460      56461      56462  \\\n",
       "token            ▁de    ▁vessie  ▁compacte        ▁du         ▁d        ôme   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56463      56464      56465      56466         56467  \\\n",
       "token            ▁et        ▁de        ▁la      ▁face  ▁postérieure   \n",
       "label              O          O          O          O             O   \n",
       "id_phrase  562101000  562101000  562101000  562101000     562101000   \n",
       "\n",
       "               56468      56469      56470        56471      56472      56473  \\\n",
       "token              ,       ▁une    ▁grande  ▁importance        ▁de      ▁base   \n",
       "label              O          O          O            O          O          O   \n",
       "id_phrase  562101000  562101000  562101000    562101000  562101000  562101000   \n",
       "\n",
       "               56474      56475      56476      56477      56478      56479  \\\n",
       "token             et      ▁dure        ▁en   ▁contact      ▁avec        ▁le   \n",
       "label              O          O          O          O          O          O   \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  562101000   \n",
       "\n",
       "               56480      56481      56482      56483      56484  \n",
       "token             ▁c          y         st     oscope          .  \n",
       "label              O          O          O          O          O  \n",
       "id_phrase  562101000  562101000  562101000  562101000  562101000  \n",
       "\n",
       "[3 rows x 954134 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from numpy import True_\n",
    "from scripts.doccano_to_camembert import JsonlDoccano as jd\n",
    "file = jd(\"\"\"input_data/exported_doccano.jsonl\"\"\", camembert=True)\n",
    "#file.run_export(\"scripts/data/tokens_v5.parquet\")\n",
    "df = file.df_augmented\n",
    "df.label = df.label.apply(lambda x: transform_label(x)) # Deleting B- I- tags\n",
    "df.T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prétraitement et tokenization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id_phrase</th>\n",
       "      <th>tokens</th>\n",
       "      <th>ner_tags</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1000</td>\n",
       "      <td>[▁Un, ▁homme, ▁âgé, ▁de, ▁77, ▁ans, ▁présentai...</td>\n",
       "      <td>[O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2000</td>\n",
       "      <td>[▁Mr, ., ▁R, ,, R, ▁âgé, ▁de, ▁53, ▁ans, ▁sans...</td>\n",
       "      <td>[O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2001</td>\n",
       "      <td>[▁Le, ▁patient, ▁a, ▁été, ▁revu, ▁1, ▁mois, ▁a...</td>\n",
       "      <td>[O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id_phrase                                             tokens  \\\n",
       "0       1000  [▁Un, ▁homme, ▁âgé, ▁de, ▁77, ▁ans, ▁présentai...   \n",
       "1       2000  [▁Mr, ., ▁R, ,, R, ▁âgé, ▁de, ▁53, ▁ans, ▁sans...   \n",
       "2       2001  [▁Le, ▁patient, ▁a, ▁été, ▁revu, ▁1, ▁mois, ▁a...   \n",
       "\n",
       "                                            ner_tags  \n",
       "0  [O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...  \n",
       "1  [O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...  \n",
       "2  [O, O, O, O, O, O, O, O, O, O, O, O, O, O, O, ...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.groupby('id_phrase').agg(lambda x: list(x)).reset_index()\n",
    "df.columns = [\"id_phrase\", \"tokens\", \"ner_tags\"]\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = truncate(df) # Toutes les listes de plus de 128 éléments sont dubdivisées en plusieurs listes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokens = df[\"tokens\"].to_list()\n",
    "ner_tags = df[\"ner_tags\"].to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Train test split\n",
    "train_tokens, val_tokens, train_tags, val_tags = train_test_split(tokens, ner_tags, test_size=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "unique_tags = ['O', 'Dosage', 'Drug', \n",
    "'Duration', 'Form', 'Frequency', 'Route', 'Treatment']\n",
    "\n",
    "tag2id = {tag: id for id, tag in enumerate(unique_tags)}\n",
    "id2tag = {id: tag for tag, id in tag2id.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"#unique_tags = set(tag for doc in ner_tags for tag in doc)\\nunique_tags = ['O', 'B-Dosage', 'I-Dosage', 'B-Drug', 'I-Drug', \\n'B-Duration', 'I-Duration', 'B-Form', 'I-Form', 'B-Frequency', \\n'I-Frequency', 'B-Route', 'I-Route', 'B-Treatment', 'I-Treatment']\\n\\ntag2id = {tag: id for id, tag in enumerate(unique_tags)}\\nid2tag = {id: tag for tag, id in tag2id.items()}\\n\""
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#To keep BIO tags\n",
    "\"\"\"#unique_tags = set(tag for doc in ner_tags for tag in doc)\n",
    "unique_tags = ['O', 'B-Dosage', 'I-Dosage', 'B-Drug', 'I-Drug', \n",
    "'B-Duration', 'I-Duration', 'B-Form', 'I-Form', 'B-Frequency', \n",
    "'I-Frequency', 'B-Route', 'I-Route', 'B-Treatment', 'I-Treatment']\n",
    "\n",
    "tag2id = {tag: id for id, tag in enumerate(unique_tags)}\n",
    "id2tag = {id: tag for tag, id in tag2id.items()}\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Tokenization\n",
    "tokenizer = CamembertTokenizerFast.from_pretrained('camembert-base')\n",
    "train_encodings = tokenizer(train_tokens, is_split_into_words=True, return_offsets_mapping=True, padding='max_length',\n",
    "                            truncation='longest_first',\n",
    "                            max_length=TOKENS_NUMBER)\n",
    "val_encodings = tokenizer(val_tokens, is_split_into_words=True, return_offsets_mapping=True, padding='max_length',\n",
    "                            truncation='longest_first',\n",
    "                            max_length=TOKENS_NUMBER)\n",
    "\n",
    "train_labels = encode_tags(train_tags, train_encodings, tag2id)\n",
    "val_labels = encode_tags(val_tags, val_encodings, tag2id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_encodings.pop(\"offset_mapping\") \n",
    "val_encodings.pop(\"offset_mapping\")\n",
    "\n",
    "train_dataset = OurDataset(train_encodings, train_labels)\n",
    "val_dataset = OurDataset(val_encodings, val_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at camembert-base were not used when initializing CamembertForTokenClassification: ['lm_head.decoder.weight', 'lm_head.bias', 'lm_head.dense.weight', 'lm_head.layer_norm.bias', 'lm_head.dense.bias', 'lm_head.layer_norm.weight']\n",
      "- This IS expected if you are initializing CamembertForTokenClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing CamembertForTokenClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "Some weights of CamembertForTokenClassification were not initialized from the model checkpoint at camembert-base and are newly initialized: ['classifier.weight', 'classifier.bias']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "model = CamembertForTokenClassification.from_pretrained('camembert-base', \n",
    "                                                        num_labels=len(unique_tags), \n",
    "                                                        id2label=id2tag, \n",
    "                                                        label2id=tag2id).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 20/20 [38:58<00:00, 116.95s/it]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "CamembertForTokenClassification(\n",
       "  (roberta): CamembertModel(\n",
       "    (embeddings): CamembertEmbeddings(\n",
       "      (word_embeddings): Embedding(32005, 768, padding_idx=1)\n",
       "      (position_embeddings): Embedding(514, 768, padding_idx=1)\n",
       "      (token_type_embeddings): Embedding(1, 768)\n",
       "      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      (dropout): Dropout(p=0.1, inplace=False)\n",
       "    )\n",
       "    (encoder): CamembertEncoder(\n",
       "      (layer): ModuleList(\n",
       "        (0): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (1): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (2): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (3): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (4): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (5): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (6): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (7): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (8): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (9): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (10): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (11): CamembertLayer(\n",
       "          (attention): CamembertAttention(\n",
       "            (self): CamembertSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): CamembertSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): CamembertIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): CamembertOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (dropout): Dropout(p=0.1, inplace=False)\n",
       "  (classifier): Linear(in_features=768, out_features=8, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True)\n",
    "\n",
    "optim = AdamW(model.parameters(), lr=5e-5)\n",
    "\n",
    "for epoch in tqdm(range(20)):\n",
    "    for batch in train_loader:\n",
    "        optim.zero_grad()\n",
    "        input_ids = batch['input_ids'].to(device)\n",
    "        attention_mask = batch['attention_mask'].to(device)\n",
    "        labels = batch['labels'].to(device)\n",
    "        outputs = model(input_ids, attention_mask=attention_mask, labels=labels)\n",
    "        loss = outputs[0]\n",
    "        loss.backward()\n",
    "        optim.step()\n",
    "\n",
    "model.eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model, 'model_v1.pb')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sur l'ensemble des vrais labels différents de O et des labels prédits différents de O, 29.5% ont le bon label prédit\n"
     ]
    }
   ],
   "source": [
    "model = torch.load('model_v1.pb')\n",
    "\n",
    "input_ids, attention_masks = val_dataset.encodings['input_ids'], val_dataset.encodings['attention_mask']\n",
    "device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')\n",
    "flat_pred = []\n",
    "with torch.no_grad():\n",
    "    input_ids, attention_masks = torch.tensor(input_ids), torch.tensor(attention_masks)\n",
    "\n",
    "    outputs =  model(input_ids.to(device), token_type_ids=None, attention_mask=attention_masks.to(device))\n",
    "\n",
    "logits = outputs['logits'].detach().cpu().numpy()\n",
    "pred = np.argmax(logits, axis=2).flatten()\n",
    "true_labels = np.array(val_labels).flatten()\n",
    "\n",
    "liste_results = []\n",
    "for i in range(len(pred)):\n",
    "    if id2tag[pred[i]] != 'O' or (true_labels[i] not in [-100, 0]):\n",
    "        liste_results.append(pred[i]==true_labels[i])\n",
    "print( \"Sur l'ensemble des vrais labels différents de O et des labels prédits différents de O, \"+\"{:.1f}\".format(np.mean(liste_results)*100)+\"% ont le bon label prédit\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.6 64-bit",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.6"
  },
  "vscode": {
   "interpreter": {
    "hash": "504a0bd1ff6cd9b5c9ceefbc2d92d99b24c0c4c6fcdb7d66189c03b11e5237d1"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}