few optimizations

cbernet · Jun 15, 2018 · 8a996ec · 8a996ec
1 parent 3439842
commit 8a996ec
Show file tree

Hide file tree

Showing 7 changed files with 159 additions and 55 deletions.
diff --git a/data b/data
@@ -0,0 +1 @@
+/data/cbernet/DeepLearning/recnn/data
diff --git a/model.pickle b/model.pickle
@@ -0,0 +1,5 @@
+((lp0
+(lp1
+(lp2
+tp3
+.
diff --git a/notebooks/04-jet-study.ipynb b/notebooks/04-jet-study.ipynb
diff --git a/notebooks/test.png b/notebooks/test.png
diff --git a/notebooks/test2.png b/notebooks/test2.png
diff --git a/recnn/preprocessing.py b/recnn/preprocessing.py
@@ -135,7 +135,7 @@ def randomize(jet):
     leaves = np.where(jet["tree"][:, 0] == -1)[0]
     nodes = [n for n in leaves]
     content = [jet["content"][n] for n in nodes]
-    nodes = [i for i in range(len(nodes))]
+    nodes = range(len(nodes))
     tree = [[-1, -1] for n in nodes]
     pool = [n for n in nodes]
     next_id = len(nodes)

diff --git a/train.py b/train.py
@@ -13,6 +13,9 @@
 from recnn.preprocessing import rewrite_content
 from recnn.preprocessing import permute_by_pt
 from recnn.preprocessing import extract
+from recnn.preprocessing import multithreadmap
+
+
 from recnn.recnn import log_loss
 from recnn.recnn import adam
 from recnn.recnn import grnn_init_simple
@@ -83,7 +86,7 @@ def train(filename_train,
 
     # Preprocessing
     logging.info("Preprocessing...")
-    X = [extract(permute_by_pt(rewrite_content(jet))) for jet in X]
+    X = multithreadmap(extract,multithreadmap(permute_by_pt,multithreadmap(rewrite_content,X)))
     tf = RobustScaler().fit(np.vstack([jet["content"] for jet in X]))
 
     for jet in X: