avoid numpy warning in split

Summary:
avoid creating a numpy array of random things just to split it: this can now generate a warning e.g. if the list contains lists of varying lengths. There might also be a performance win here, and we could do more of the same if we care about that.

(The vanilla way to avoid the new warning is to replace `np.split(a,` with `np.split(np.array(a, dtype=object), ` btw.)

Reviewed By: shapovalov

Differential Revision: D40209308

fbshipit-source-id: daae33a23ceb444e8e7241f72ce1525593e2f239

pytorch3d/implicitron/dataset/types.py

@@ -225,8 +225,8 @@ def _dataclass_list_from_dict_list(dlist, typeannot):
         assert indices[-1] == len(all_keys_res)
 
         keys = np.split(list(all_keys_res), indices[:-1])
-        vals = np.split(list(all_vals_res), indices[:-1])
-        return [cls(zip(k, v)) for k, v in zip(keys, vals)]
+        all_vals_res_iter = iter(all_vals_res)
+        return [cls(zip(k, all_vals_res_iter)) for k in keys]
     elif not dataclasses.is_dataclass(typeannot):
         return dlist