Merge pull request #86 from WuHaobo/master

polish mixup and eval

tools/eval.py

@@ -74,7 +74,7 @@ def main(args):
 
     compiled_valid_prog = program.compile(config, valid_prog)
     program.run(valid_dataloader, exe, compiled_valid_prog, valid_fetchs, -1,
-                'valid')
+                'eval')
 
 
 if __name__ == '__main__':

tools/program.py

@@ -385,19 +385,19 @@ def run(dataloader, exe, program, fetchs, epoch=0, mode='train'):
             metric_list[i].update(m[0], len(batch[0]))
         fetchs_str = ''.join([str(m.value) + ' '
                               for m in metric_list] + [batch_time.value])
-        if epoch != -1:
+        if mode == 'eval':
+            logger.info("{:s} step:{:<4d} {:s}s".format(mode, idx, fetchs_str))
+        else:
             logger.info("epoch:{:<3d} {:s} step:{:<4d} {:s}s".format(
                 epoch, mode, idx, fetchs_str))
-        else:
-            logger.info("{:s} step:{:<4d} {:s}s".format(mode, idx, fetchs_str))
 
     end_str = ''.join([str(m.mean) + ' '
                        for m in metric_list] + [batch_time.total])
-    if epoch != -1:
-        logger.info("END epoch:{:<3d} {:s} {:s}s".format(epoch, mode, end_str))
-    else:
+    if mode == 'eval':
         logger.info("END {:s} {:s}s".format(mode, end_str))
+    else:
+        logger.info("END epoch:{:<3d} {:s} {:s}s".format(epoch, mode, end_str))
 
-    # save the best model
-    top1_acc = fetchs["top1"][1].avg
-    return top1_acc
+    # return top1_acc in order to save the best model
+    if mode == 'valid':
+        return fetchs["top1"][1].avg