Use percentage to show the results

README.md

@@ -34,7 +34,7 @@ The following table demonstrates the testing accuracy (%) comparisons on differe
 
 ### Semi-supervised Setting Results
 The following table demonstrates the testing accuracy (%) comparisons on different backbones and layers w and w/o DropEdge.
-<table><tr><th rowspan="2">Dataset</th><th rowspan="2">Method</th><th colspan="2">2 layers</th><th colspan="2">4 layers</th><th colspan="2">8 layers</th><th colspan="2">16 layers</th><th colspan="2">32 layers</th><th colspan="2">64 layers</th></tr><tr><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td></tr><tr><td rowspan="4">Cora</td><td>GCN</td><td>0.811</td><td>0.828</td><td>0.804</td><td>0.82</td><td>0.695</td><td>0.758</td><td>0.649</td><td>0.757</td><td>0.603</td><td>0.625</td><td>0.287</td><td>0.495</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>0.788</td><td>0.833</td><td>0.756</td><td>0.828</td><td>0.722</td><td>0.827</td><td>0.766</td><td>0.811</td><td>0.611</td><td>0.789</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>0.802</td><td>0.833</td><td>0.807</td><td>0.826</td><td>0.802</td><td>0.83</td><td>0.811</td><td>0.825</td><td>0.715</td><td>0.832</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>0.776</td><td>0.829</td><td>0.765</td><td>0.825</td><td>0.817</td><td>0.831</td><td>0.817</td><td>0.831</td><td>0.8</td><td>0.835</td></tr><tr><td rowspan="4">Citeseer</td><td>GCN</td><td>0.708</td><td>0.723</td><td>0.676</td><td>0.706</td><td>0.302</td><td>0.614</td><td>0.183</td><td>0.572</td><td>0.25</td><td>0.416</td><td>0.2</td><td>0.344</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>0.705</td><td>0.722</td><td>0.65</td><td>0.716</td><td>0.665</td><td>0.701</td><td>0.626</td><td>0.7</td><td>0.221</td><td>0.651</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>0.687</td><td>0.726</td><td>0.677</td><td>0.718</td><td>0.698</td><td>0.726</td><td>0.682</td><td>0.708</td><td>0.634</td><td>0.722</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>0.693</td><td>0.727</td><td>0.684</td><td>0.714</td><td>0.702</td><td>0.725</td><td>0.68</td><td>0.726</td><td>0.675</td><td>0.71</td></tr><tr><td rowspan="4">Pubmed</td><td>GCN</td><td>0.79</td><td>0.796</td><td>0.765</td><td>0.794</td><td>0.612</td><td>0.781</td><td>0.409</td><td>0.785</td><td>0.224</td><td>0.77</td><td>0.353</td><td>0.615</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>0.786</td><td>0.788</td><td>0.781</td><td>0.789</td><td>0.755</td><td>0.78</td><td>0.679</td><td>0.782</td><td>0.669</td><td>0.769</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>0.78</td><td>0.787</td><td>0.781</td><td>0.787</td><td>0.726</td><td>0.791</td><td>0.724</td><td>0.792</td><td>0.745</td><td>0.789</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>0.777</td><td>0.795</td><td>0.779</td><td>0.786</td><td>0.749</td><td>0.79</td><td>OOM</td><td>OOM</td><td>OOM</td><td>OOM</td></tr></table>
+<table><tr><th rowspan="2">Dataset</th><th rowspan="2">Method</th><th colspan="2">2 layers</th><th colspan="2">4 laysers</th><th colspan="2">8 layers</th><th colspan="2">16 layers</th><th colspan="2">32 layers</th><th colspan="2">64 layers</th></tr><tr><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td><td>Orignal</td><td>DropEdge</td></tr><tr><td rowspan="4">Cora</td><td>GCN</td><td>81.10</td><td>82.80</td><td>80.40</td><td>82.00</td><td>69.50</td><td>75.80</td><td>64.90</td><td>75.70</td><td>60.30</td><td>62.50</td><td>28.70</td><td>49.50</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>78.80</td><td>83.30</td><td>75.60</td><td>82.80</td><td>72.20</td><td>82.70</td><td>76.60</td><td>81.10</td><td>61.10</td><td>78.90</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>80.20</td><td>83.30</td><td>80.70</td><td>82.60</td><td>80.20</td><td>83.00</td><td>81.10</td><td>82.50</td><td>71.50</td><td>83.20</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>77.60</td><td>82.90</td><td>76.50</td><td>82.50</td><td>81.70</td><td>83.10</td><td>81.70</td><td>83.10</td><td>80.00</td><td>83.50</td></tr><tr><td rowspan="4">Citeseer</td><td>GCN</td><td>70.80</td><td>72.30</td><td>67.60</td><td>70.60</td><td>30.20</td><td>61.40</td><td>18.30</td><td>57.20</td><td>25.00</td><td>41.60</td><td>20.00</td><td>34.40</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>70.50</td><td>72.20</td><td>65.00</td><td>71.60</td><td>66.50</td><td>70.10</td><td>62.60</td><td>70.00</td><td>22.10</td><td>65.10</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>68.70</td><td>72.60</td><td>67.70</td><td>71.80</td><td>69.80</td><td>72.60</td><td>68.20</td><td>70.80</td><td>63.40</td><td>72.20</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>69.30</td><td>72.70</td><td>68.40</td><td>71.40</td><td>70.20</td><td>72.50</td><td>68.00</td><td>72.60</td><td>67.50</td><td>71.00</td></tr><tr><td rowspan="4">Pubmed</td><td>GCN</td><td>79.00</td><td>79.60</td><td>76.50</td><td>79.40</td><td>61.20</td><td>78.10</td><td>40.90</td><td>78.50</td><td>22.40</td><td>77.00</td><td>35.30</td><td>61.50</td></tr><tr><td>ResGCN</td><td>-</td><td>-</td><td>78.60</td><td>78.80</td><td>78.10</td><td>78.90</td><td>75.50</td><td>78.00</td><td>67.90</td><td>78.20</td><td>66.90</td><td>76.90</td></tr><tr><td>JKNet</td><td>-</td><td>-</td><td>78.00</td><td>78.70</td><td>78.10</td><td>78.70</td><td>72.60</td><td>79.10</td><td>72.40</td><td>79.20</td><td>74.50</td><td>78.90</td></tr><tr><td>IncepGCN</td><td>-</td><td>-</td><td>77.70</td><td>79.50</td><td>77.90</td><td>78.60</td><td>74.90</td><td>79.00</td><td>OOM</td><td>OOM</td><td>OOM</td><td>OOM</td></tr></table>
 
 ## Change Log
  * 2019-10-11: Support both full-supervised and semi-supervised task setting for `Cora`, `Citeseer` and `Pubmed`. See `--task_type` option.