Update tensorboard logging to plot test time details (#745)

Summary:
Pull Request resolved: #745

It is helpful to know the memory consumption and speed during test phases. Removed the epoch duration since it can be derived from the images / sec numbers

Reviewed By: lauragustafson

Differential Revision: D28308240

fbshipit-source-id: 08a7b2bed4a98528f94a88c90fcc5a2221a0e8a3

classy_vision/hooks/tensorboard_plot_hook.py

@@ -164,16 +164,16 @@ def on_phase_end(self, task) -> None:
                     f"Parameters/{name}", parameter, global_step=phase_type_idx
                 )
 
-        if torch.cuda.is_available() and task.train:
+        if torch.cuda.is_available():
             self.tb_writer.add_scalar(
-                "Memory/peak_allocated",
+                f"Memory/{phase_type}/peak_allocated",
                 torch.cuda.max_memory_allocated(),
                 global_step=phase_type_idx,
             )
 
         loss_avg = sum(task.losses) / batches
 
-        loss_key = "Losses/{phase_type}".format(phase_type=task.phase_type)
+        loss_key = f"Losses/{phase_type}"
         self.tb_writer.add_scalar(loss_key, loss_avg, global_step=phase_type_idx)
 
         # plot meters which return a dict

classy_vision/tasks/classification_task.py

@@ -1296,7 +1296,7 @@ def on_phase_start(self):
         self.phase_start_time_train = time.perf_counter()
 
     def on_phase_end(self):
-        self.log_phase_end("train")
+        self.log_phase_end(self.phase_type)
 
         if self.train:
             self.optimizer.on_epoch(where=self.where)
@@ -1315,7 +1315,7 @@ def on_phase_end(self):
             hook.on_phase_end(self)
         self.perf_log = []
 
-        self.log_phase_end("total")
+        self.log_phase_end(f"{self.phase_type}_total")
 
         if hasattr(self.datasets[self.phase_type], "on_phase_end"):
             self.datasets[self.phase_type].on_phase_end()
@@ -1325,12 +1325,9 @@ def on_end(self):
             hook.on_end(self)
 
     def log_phase_end(self, tag):
-        if not self.train:
-            return
-
         start_time = (
             self.phase_start_time_train
-            if tag == "train"
+            if tag == self.phase_type
             else self.phase_start_time_total
         )
         phase_duration = time.perf_counter() - start_time
@@ -1341,7 +1338,6 @@ def log_phase_end(self, tag):
             {
                 "tag": tag,
                 "phase_idx": self.train_phase_idx,
-                "epoch_duration": phase_duration,
                 "im_per_sec": im_per_sec,
             }
         )