Add ce for icnet.

fluid/icnet/.run_ce.sh

+#!/bin/bash
+
+# This file is only used for continuous evaluation.
+rm -rf ./ck
+mkdir ck
+python train.py --use_gpu=True --checkpoint_path="./ck"; python eval.py  --model_path="./ck/100" | python _ce.py

fluid/icnet/_ce.py

+# this file is only used for continuous evaluation test!
+
+import os
+import sys
+sys.path.append(os.environ['ceroot'])
+from kpi import CostKpi, DurationKpi, AccKpi
+
+# NOTE kpi.py should shared in models in some way!!!!
+
+train_cost_kpi = CostKpi('train_cost', 0.02, actived=True)
+test_acc_kpi = AccKpi('test_acc', 0.005, actived=True)
+train_duration_kpi = DurationKpi('train_duration', 0.06, actived=True)
+
+tracking_kpis = [
+    train_cost_kpi,
+    test_acc_kpi,
+    train_duration_kpi,
+]
+
+
+def parse_log(log):
+    '''
+    This method should be implemented by model developers.
+
+    The suggestion:
+
+    each line in the log should be key, value, for example:
+
+    "
+    train_cost\t1.0
+    test_cost\t1.0
+    train_cost\t1.0
+    train_cost\t1.0
+    train_acc\t1.2
+    "
+    '''
+    for line in log.split('\n'):
+        fs = line.strip().split('\t')
+        print(fs)
+        if len(fs) == 3 and fs[0] == 'kpis':
+            kpi_name = fs[1]
+            kpi_value = float(fs[2])
+            yield kpi_name, kpi_value
+
+
+def log_to_ce(log):
+    kpi_tracker = {}
+    for kpi in tracking_kpis:
+        kpi_tracker[kpi.name] = kpi
+
+    for (kpi_name, kpi_value) in parse_log(log):
+        print(kpi_name, kpi_value)
+        kpi_tracker[kpi_name].add_record(kpi_value)
+        kpi_tracker[kpi_name].persist()
+
+
+if __name__ == '__main__':
+    log = sys.stdin.read()
+    log_to_ce(log)

fluid/icnet/eval.py

@@ -84,6 +84,7 @@ def eval(args):
         sys.stdout.flush()
     iou = cal_mean_iou(out_wrong, out_right)
     print "\nmean iou: %.3f" % iou
+    print "kpis	test_acc	%f" % iou
 
 
 def main():

fluid/icnet/train.py

@@ -11,6 +11,10 @@ from utils import add_arguments, print_arguments, get_feeder_data
 from paddle.fluid.layers.learning_rate_scheduler import _decay_step_counter
 from paddle.fluid.initializer import init_on_cpu
 
+SEED = 90
+# random seed must set before configuring the network.
+fluid.default_startup_program().random_seed = SEED
+
 parser = argparse.ArgumentParser(description=__doc__)
 add_arg = functools.partial(add_arguments, argparser=parser)
 # yapf: disable
@@ -27,9 +31,9 @@ LAMBDA2 = 0.4
 LAMBDA3 = 1.0
 LEARNING_RATE = 0.003
 POWER = 0.9
-LOG_PERIOD = 1
-CHECKPOINT_PERIOD = 1000
-TOTAL_STEP = 60000
+LOG_PERIOD = 100
+CHECKPOINT_PERIOD = 100
+TOTAL_STEP = 100
 
 no_grad_set = []
 
@@ -97,10 +101,13 @@ def train(args):
     sub124_loss = 0.
     train_reader = cityscape.train(
         args.batch_size, flip=args.random_mirror, scaling=args.random_scaling)
+    start_time = time.time()
     while True:
         # train a pass
         for data in train_reader():
             if iter_id > TOTAL_STEP:
+                end_time = time.time()
+                print "kpis	train_duration	%f" % (end_time - start_time)
                 return
             iter_id += 1
             results = exe.run(
@@ -115,6 +122,8 @@ def train(args):
                 print "Iter[%d]; train loss: %.3f; sub4_loss: %.3f; sub24_loss: %.3f; sub124_loss: %.3f" % (
                     iter_id, t_loss / LOG_PERIOD, sub4_loss / LOG_PERIOD,
                     sub24_loss / LOG_PERIOD, sub124_loss / LOG_PERIOD)
+                print "kpis	train_cost	%f" % (t_loss / LOG_PERIOD)
+
                 t_loss = 0.
                 sub4_loss = 0.
                 sub24_loss = 0.