Logging includes %done and ETA via a util function

parlai/core/utils.py

@@ -101,6 +101,8 @@ def load_cands(path, lines_have_ids = False, cands_are_replies = False):
     return cands
 
 
+
+
 class Predictor(object):
     """Provides functionality for setting up a running version of a model and
     requesting predictions from that model on live data.
@@ -175,6 +177,34 @@ def time(self):
         return self.total
 
 
+class TimeLogger():
+    def __init__(self):
+        self.timer = Timer()
+        self.tot_time = 0
+
+    def total_time(self):
+        return self.tot_time
+
+    def time(self):
+        return self.timer.time()
+
+    def log(self, done, total, report={}):
+        self.tot_time += self.timer.time()
+        self.timer.reset()
+        log = {}
+        log['total'] = done
+        if total > 0:
+            log['%done'] = done / total
+            if log["%done"] > 0:
+                log['time_left'] = str(int(self.tot_time / log['%done'] - self.tot_time)) + 's'
+            z = '%.2f' % ( 100*log['%done'])
+            log['%done'] = str(z) + '%'  
+        for k, v in report.items():
+            if k not in log:
+                log[k] = v
+        text = str(int(self.tot_time)) + "s elapsed: " + str(log)
+        return text, log
+
 class AttrDict(dict):
     """Helper class to have a dict-like object with dot access.
 

parlai/core/worlds.py

@@ -644,6 +644,9 @@ def num_examples(self):
     def num_episodes(self):
         return self.world.num_episodes()
 
+    def get_total_exs(self):
+        return self.world.get_total_exs()
+
     def getID(self):
         return self.world.getID()
 
@@ -833,6 +836,9 @@ def num_examples(self):
     def num_episodes(self):
         return self.inner_world.num_episodes()
 
+    def get_total_exs(self):
+        return self.inner_world.get_total_exs()
+
     def get_total_epochs(self):
         """Return total amount of epochs on which the world has trained."""
         if self.max_exs is None:

parlai/scripts/build_dict.py

@@ -8,9 +8,10 @@
 from parlai.core.dict import DictionaryAgent
 from parlai.core.params import ParlaiParser, str2class
 from parlai.core.worlds import create_task
+from parlai.core.utils import TimeLogger
 import copy
 import os
-
+import sys
 
 def setup_args(parser=None):
     if parser is None:
@@ -22,6 +23,7 @@ def setup_args(parser=None):
         help='Include validation set in dictionary building for task.')
     dict_loop.add_argument('--dict-include-test', default=False, type='bool',
         help='Include test set in dictionary building for task.')
+    dict_loop.add_argument('-ltim', '--log-every-n-secs', type=float, default=2)
     partial, _ = parser.parse_known_args(nohelp=True)
     if vars(partial).get('dict_class'):
         str2class(vars(partial).get('dict_class')).add_cmdline_args(parser)
@@ -37,7 +39,6 @@ def build_dict(opt, skip_if_built=False):
         print('Tried to build dictionary but `--dict-file` is not set. Set ' +
               'this param so the dictionary can be saved.')
         return
-    print('[ setting up dictionary. ]')
 
     if skip_if_built and os.path.isfile(opt['dict_file']):
         # Dictionary already built, skip all loading or setup
@@ -78,13 +79,25 @@ def build_dict(opt, skip_if_built=False):
         ordered_opt['datatype'] = dt
         world_dict = create_task(ordered_opt, dictionary)
         # pass examples to dictionary
+        print('[ running dictionary over data.. ]')
+        log_every_n_secs = opt.get('log_every_n_secs', -1)
+        if log_every_n_secs <= 0:
+            log_every_n_secs = float('inf')
+        log_time = TimeLogger()
         while not world_dict.epoch_done():
             cnt += 1
             if cnt > opt['dict_maxexs'] and opt['dict_maxexs'] > 0:
                 print('Processed {} exs, moving on.'.format(opt['dict_maxexs']))
                 # don't wait too long...
                 break
             world_dict.parley()
+            if log_time.time() > log_every_n_secs:
+                sys.stdout.write('\r')
+                text, _log = log_time.log(cnt, max(opt.get('dict_maxexs',0),
+                                                   world_dict.num_examples()))
+                sys.stdout.write(text)
+                sys.stdout.flush()
+
     dictionary.save(opt['dict_file'], sort=True)
     print('[ dictionary built with {} tokens ]'.format(len(dictionary)))
     return dictionary

parlai/scripts/detect_offensive_language.py

@@ -13,7 +13,7 @@
 from parlai.core.agents import create_agent
 from parlai.core.worlds import create_task
 from parlai.agents.repeat_label.repeat_label import RepeatLabelAgent
-from parlai.core.utils import Timer, OffensiveLanguageDetector
+from parlai.core.utils import OffensiveLanguageDetector, TimeLogger
 
 import random
 
@@ -50,8 +50,7 @@ def detect(opt, printargs=None, print_parser=None):
     log_every_n_secs = opt.get('log_every_n_secs', -1)
     if log_every_n_secs <= 0:
         log_every_n_secs = float('inf')
-    log_time = Timer()
-    tot_time = 0
+    log_time = TimeLogger()
 
     # Show some example dialogs:
     cnt = 0
@@ -71,17 +70,11 @@ def detect(opt, printargs=None, print_parser=None):
                 print(world.display() + "\n~~")
             cnt += 1
         if log_time.time() > log_every_n_secs:
-            tot_time += log_time.time()
             report = world.report()
-            log = {'total': report['total']}
-            log['done'] = report['total'] / world.num_examples()
-            if log['done'] > 0:
-                log['eta'] = int(tot_time / log['done'] - tot_time)
-            z = '%.2f' % ( 100*log['done'])
-            log['done'] = str(z) + '%'
-            log['offenses'] = cnt
-            print(str(int(tot_time)) + "s elapsed: " + str(log))
-            log_time.reset()
+            log = { 'offenses': cnt }
+            text, log = log_time.log(report['total'], world.num_examples(), log)
+            print(text)
+
     if world.epoch_done():
         print("EPOCH DONE")
     print(str(cnt) + " offensive messages found out of " +

parlai/scripts/eval_model.py

@@ -14,7 +14,7 @@
 from parlai.core.params import ParlaiParser
 from parlai.core.agents import create_agent
 from parlai.core.worlds import create_task
-from parlai.core.utils import Timer
+from parlai.core.utils import TimeLogger
 
 import random
 import os
@@ -63,8 +63,7 @@ def eval_model(opt, printargs=None, print_parser=None):
     log_every_n_secs = opt.get('log_every_n_secs', -1)
     if log_every_n_secs <= 0:
         log_every_n_secs = float('inf')
-    log_time = Timer()
-    tot_time = 0
+    log_time = TimeLogger()
 
     # Show some example dialogs:
     cnt = 0
@@ -74,9 +73,9 @@ def eval_model(opt, printargs=None, print_parser=None):
         if opt['display_examples']:
             print(world.display() + "\n~~")
         if log_time.time() > log_every_n_secs:
-            tot_time += log_time.time()
-            print(str(int(tot_time)) + "s elapsed: " + str(world.report()))
-            log_time.reset()
+            report = world.report()
+            text, report = log_time.log(report['total'], world.num_examples(), report)
+            print(text)
         if opt['num_examples'] > 0 and cnt >= opt['num_examples']:
             break
     if world.epoch_done():

parlai/scripts/train_model.py

@@ -236,7 +236,7 @@ def log(self):
 
         # time elapsed
         logs.append('time:{}s'.format(math.floor(self.train_time.time())))
-        logs.append('parleys:{}'.format(self.parleys))
+        logs.append('total_exs:{}'.format(self.world.get_total_exs()))
 
         if 'time_left' in train_report:
             logs.append('time_left:{}s'.format(

projects/personachat/kvmemnn/kvmemnn.py

@@ -710,10 +710,10 @@ def valid(obs):
         if ys is None:
             # only build candidates in eval mode.
             for o in observations:
-                if 'label_candidates' in o:
+                if 'label_candidates' in o and o['label_candidates'] is not None:
                     cs = []
                     ct = []
-                    for c in o['label_candidates'] and o['label_candidates'] is not None:
+                    for c in o['label_candidates']:
                         cs.append(Variable(torch.LongTensor(self.parse(c)).unsqueeze(0)))
                         ct.append(c)
                     cands.append(cs)