From 3162664ad7d0aceaa2a9fcb7f89825a2dff83cf2 Mon Sep 17 00:00:00 2001
From: Qiao Longfei <qiaolongfei@baidu.com>
Date: Wed, 29 May 2019 10:24:43 +0800
Subject: [PATCH] fix some problem

---
 PaddleRec/ctr/dataset/cluster_train.sh      | 8 ++++----
 PaddleRec/ctr/dataset/ctr_dataset_reader.py | 2 +-
 PaddleRec/ctr/dataset/train_dataset.py      | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/PaddleRec/ctr/dataset/cluster_train.sh b/PaddleRec/ctr/dataset/cluster_train.sh
index 9a994a45..00dd1b68 100644
--- a/PaddleRec/ctr/dataset/cluster_train.sh
+++ b/PaddleRec/ctr/dataset/cluster_train.sh
@@ -1,7 +1,7 @@
 #!/bin/bash
 
 # start pserver0
-python train.py \
+python train_dataset.py \
     --train_data_path /paddle/data/train.txt \
     --is_local 0 \
     --role pserver \
@@ -11,7 +11,7 @@ python train.py \
     > pserver0.log 2>&1 &
 
 # start pserver1
-python train.py \
+python train_dataset.py \
     --train_data_path /paddle/data/train.txt \
     --is_local 0 \
     --role pserver \
@@ -21,7 +21,7 @@ python train.py \
     > pserver1.log 2>&1 &
 
 # start trainer0
-python train.py \
+python train_dataset.py \
     --train_data_path /paddle/data/train.txt \
     --is_local 0 \
     --role trainer \
@@ -31,7 +31,7 @@ python train.py \
     > trainer0.log 2>&1 &
 
 # start trainer1
-python train.py \
+python train_dataset.py \
     --train_data_path /paddle/data/train.txt \
     --is_local 0 \
     --role trainer \
diff --git a/PaddleRec/ctr/dataset/ctr_dataset_reader.py b/PaddleRec/ctr/dataset/ctr_dataset_reader.py
index 6ea31771..dc75406c 100644
--- a/PaddleRec/ctr/dataset/ctr_dataset_reader.py
+++ b/PaddleRec/ctr/dataset/ctr_dataset_reader.py
@@ -60,7 +60,7 @@ class CriteoDataset(data_generator.MultiSlotDataGenerator):
                     dense_feature.append((float(features[idx]) - self.cont_min_[idx - 1]) / self.cont_diff_[idx - 1])
             ret_result.append(("dense_feature", dense_feature))
             for idx in self.categorical_range_:
-                ret_result.append((str(idx - 13), [hash(str(idx) + features[idx]) % self.hash_dim_]))
+                ret_result.append(("C" + str(idx - 13), [hash(str(idx) + features[idx]) % self.hash_dim_]))
             ret_result.append(("label", [int(features[0])]))
 
             yield tuple(ret_result)
diff --git a/PaddleRec/ctr/dataset/train_dataset.py b/PaddleRec/ctr/dataset/train_dataset.py
index da49ffe4..22605736 100644
--- a/PaddleRec/ctr/dataset/train_dataset.py
+++ b/PaddleRec/ctr/dataset/train_dataset.py
@@ -213,7 +213,7 @@ def train(args):
         dataset.set_batch_size(128)
         dataset.set_use_var(words)
         pipe_command = 'python ctr_dataset_reader.py %d %d %d %d' \
-                       % args.sparse_feature_dim, args.trainer_id, 1, args.trainer_num
+                       % (args.sparse_feature_dim, args.trainer_id, 1, args.trainers)
         dataset.set_pipe_command(pipe_command)
 
         dataset.set_filelist(filelist)
-- 
GitLab