add support for vdl (#978)

* add support for vdl

* fix vdl eval

ppcls/engine/trainer.py

@@ -26,6 +26,7 @@ import argparse
 import paddle
 import paddle.nn as nn
 import paddle.distributed as dist
+from visualdl import LogWriter
 
 from ppcls.utils.check import check_gpu
 from ppcls.utils.misc import AverageMeter
@@ -83,8 +84,7 @@ class Trainer(object):
             self.model = paddle.DataParallel(self.model)
 
         self.vdl_writer = None
-        if self.config['Global']['use_visualdl']:
-            from visualdl import LogWriter
+        if self.config['Global']['use_visualdl'] and mode == "train":
             vdl_writer_path = os.path.join(self.output_dir, "vdl")
             if not os.path.exists(vdl_writer_path):
                 os.makedirs(vdl_writer_path)
@@ -219,6 +219,18 @@ class Trainer(object):
                             "epochs"], iter_id,
                                len(self.train_dataloader), lr_msg, metric_msg,
                                time_msg, ips_msg, eta_msg))
+
+                    logger.scaler(
+                        name="lr",
+                        value=lr_sch.get_lr(),
+                        step=global_step,
+                        writer=self.vdl_writer)
+                    for key in output_info:
+                        logger.scaler(
+                            name="train_{}".format(key),
+                            value=output_info[key].avg,
+                            step=global_step,
+                            writer=self.vdl_writer)
                 tic = time.time()
 
             metric_msg = ", ".join([
@@ -246,6 +258,12 @@ class Trainer(object):
                         prefix="best_model")
                 logger.info("[Eval][Epoch {}][best metric: {}]".format(
                     epoch_id, best_metric["metric"]))
+                logger.scaler(
+                    name="eval_acc",
+                    value=acc,
+                    step=epoch_id,
+                    writer=self.vdl_writer)
+
                 self.model.train()
 
             # save model
@@ -266,6 +284,9 @@ class Trainer(object):
                     model_name=self.config["Arch"]["name"],
                     prefix="latest")
 
+        if self.vdl_writer is not None:
+            self.vdl_writer.close()
+
     def build_avg_metrics(self, info_dict):
         return {key: AverageMeter(key, '7.5f') for key in info_dict}
 

ppcls/utils/logger.py

@@ -102,6 +102,8 @@ def scaler(name, value, step, writer):
            visualdl --logdir ./scalar --host 0.0.0.0 --port 8830 
            to preview loss corve in real time.
     """
+    if writer is None:
+        return
     writer.add_scalar(tag=name, step=step, value=value)