fix a bug which breaks the GAN training

Summary:
The GAN training pipeline is broken since last week, with the TypeError: cannot pickle
'_io.TextIOWrapper' object (refer to f283777469 for more details)

The issue is caused by D29379832 (https://github.com/facebookresearch/d2go/commit/5509a1383c1162081e9784d79eaf0b12ebbca1fd), in which the deepcopy is failed. This diff fixes this issue by disabling the FLOPs calculation when the deepcopy is failed.

Reviewed By: ppwwyyxx

Differential Revision: D29552182

fbshipit-source-id: 80d078b3d8ca68535e0366a412668e098b04ed04

d2go/utils/flop_calculator.py

@@ -33,7 +33,13 @@ def dump_flops_info(model, inputs, output_dir, use_eval_mode=True):
     if not comm.is_main_process():
         return
     logger.info("Evaluating model's number of parameters and FLOPS")
-    model = copy.deepcopy(model)
+
+    try:
+        model = copy.deepcopy(model)
+    except Exception:
+        logger.info("Failed to deepcopy the model and skip FlopsEstimation.")
+        return
+
     if use_eval_mode:
         model.eval()
     inputs = copy.deepcopy(inputs)