Polish PointNet++ model (#4009)

* fix python3 compatable.

* update weight download link

PaddleCV/Paddle3D/PointNet++/README.md

@@ -117,7 +117,7 @@ sh download.sh
 
 **Indoor3DSemSeg 数据集:**
 
-PointNet++ 分类模型在 [Indoor3DSemSeg 数据集](https://shapenet.cs.stanford.edu/media/indoor3d_sem_seg_hdf5_data.zip)上进行训练，我们提供了数据集下载脚本：
+PointNet++ 分割模型在 [Indoor3DSemSeg 数据集](https://shapenet.cs.stanford.edu/media/indoor3d_sem_seg_hdf5_data.zip)上进行训练，我们提供了数据集下载脚本：
 
 ```
 cd dataset/Indoor3DSemSeg
@@ -212,8 +212,8 @@ sh scripts/eval_cls.sh
 
 | model | Top-1 | download |
 | :----- | :---: | :---: |
-| SSG(Single-Scale Group) | 87.4 | [model]() |
-| MSG(Multi-Scale Group)  | 89.2 | [model]() |
+| SSG(Single-Scale Group) | 87.4 | [model](https://paddlemodels.bj.bcebos.com/Paddle3D/pointnet2_ssg_cls.tar) |
+| MSG(Multi-Scale Group)  | 89.1 | [model](https://paddlemodels.bj.bcebos.com/Paddle3D/pointnet2_msg_cls.tar) |
 
 **语义分割模型:**
 
@@ -234,8 +234,8 @@ sh scripts/eval_seg.sh
 
 | model | Top-1 | download |
 | :----- | :---: | :---: |
-| SSG(Single-Scale Group) | 86.1 | [model]() |
-| MSG(Multi-Scale Group)  | 86.8 | [model]() |
+| SSG(Single-Scale Group) | 86.1 | [model](https://paddlemodels.bj.bcebos.com/Paddle3D/pointnet2_ssg_seg.tar) |
+| MSG(Multi-Scale Group)  | 86.6 | [model](https://paddlemodels.bj.bcebos.com/Paddle3D/pointnet2_msg_seg.tar) |
 
 ## 参考文献
 

PaddleCV/Paddle3D/PointNet++/data/__init__.py

@@ -14,8 +14,8 @@
 
 from . import indoor3d_reader
 from . import modelnet40_reader
-from indoor3d_reader import *
-from modelnet40_reader import *
+from .indoor3d_reader import *
+from .modelnet40_reader import *
 
 __all__ = indoor3d_reader.__all__
 __all__ += modelnet40_reader.__all__

PaddleCV/Paddle3D/PointNet++/eval_cls.py

@@ -32,6 +32,8 @@ FORMAT = '%(asctime)s-%(levelname)s: %(message)s'
 logging.basicConfig(level=logging.INFO, format=FORMAT, stream=sys.stdout)
 logger = logging.getLogger(__name__)
 
+np.random.seed(1024)
+
 
 def parse_args():
     parser = argparse.ArgumentParser("PointNet++ semantic segmentation train script")

PaddleCV/Paddle3D/PointNet++/eval_seg.py

@@ -31,6 +31,8 @@ FORMAT = '%(asctime)s-%(levelname)s: %(message)s'
 logging.basicConfig(level=logging.INFO, format=FORMAT, stream=sys.stdout)
 logger = logging.getLogger(__name__)
 
+np.random.seed(1024)
+
 
 def parse_args():
     parser = argparse.ArgumentParser("PointNet++ semantic segmentation train script")

PaddleCV/Paddle3D/PointNet++/ext_op/__init__.py

@@ -13,6 +13,6 @@
 #limitations under the License.
 
 from . import pointnet_lib
-from pointnet_lib import *
+from .pointnet_lib import *
 
 __all__ = pointnet_lib.__all__

PaddleCV/Paddle3D/PointNet++/models/pointnet2_modules.py

@@ -197,14 +197,14 @@ def pointnet_fp_module(unknown, known, unknown_feats, known_feats, mlp, bn=True,
         raise NotImplementedError("Not implement known as None currently.")
     else:
         dist, idx = three_nn(unknown, known, eps=0)
-	dist.stop_gradient = True
-	idx.stop_gradient = True
+        dist.stop_gradient = True
+        idx.stop_gradient = True
         dist = fluid.layers.sqrt(dist)
         ones = fluid.layers.fill_constant_batch_size_like(dist, dist.shape, dist.dtype, 1)
         dist_recip = ones / (dist + 1e-8); # 1.0 / dist
         norm = fluid.layers.reduce_sum(dist_recip, dim=-1, keep_dim=True)
         weight = dist_recip / norm
-	weight.stop_gradient = True
+        weight.stop_gradient = True
         interp_feats = three_interp(known_feats, weight, idx)
 
     new_features = interp_feats if unknown_feats is None else \

PaddleCV/Paddle3D/PointNet++/models/pointnet2_seg.py

@@ -98,7 +98,7 @@ class PointNet2SemSeg(object):
         self.acc1 = fluid.layers.accuracy(pred, label, k=1)
 
     def get_feeds(self):
-	return self.feed_vars
+        return self.feed_vars
 
     def get_outputs(self):
         return {"loss": self.loss, "accuracy": self.acc1}