removing warning since it isn't working

nbs/05_data.transforms.ipynb

@@ -679,7 +679,7 @@
    "outputs": [],
    "source": [
     "# export\n",
-    "def GroupedSplitter(groupkey,valid_pct=0.2,seed=None,n_tries=3,suppress_warning=False):\n",
+    "def GroupedSplitter(groupkey,valid_pct=0.2,seed=None,n_tries=3):\n",
     "    \"Splits groups of items between train/val randomly, such that val should have close to `valid_pct` of the total number of items (similar to RandomSplitter). Groups are defined by a `groupkey`, which is a callable to apply to individual items to get the groupname, or a column name if `o` is a DataFrame\"\n",
     "    def _inner(o):\n",
     "        if callable(groupkey):\n",
@@ -714,10 +714,6 @@
     "        shuffled_gk['is_valid']=([True] * valid_rows + \n",
     "                                 [False]*(len(shuffled_gk) - valid_rows))\n",
     "        split_df=ids.join(shuffled_gk.loc[:,'is_valid'],on=keycol)\n",
-    "        if ~suppress_warning and -split_goodness/desired_valid>.25:\n",
-    "            import warnings\n",
-    "            actual_split=split_df.groupby('is_valid').count()[0].to_string(header=False,index=False).replace('\\n','/').replace(' ','')\n",
-    "            warnings.warn(f'actual train/val split {actual_split} is significantly different from `valid_pct` requested, consider raising `best_split_of`, but if you have very few categories this may be unavoidable. This warning is suppressable with `suppress_warning`')      \n",
     "        return ColSplitter()(split_df)\n",
     "    return _inner"
    ]