Don't divide steps by NUM_TPUS if passed from command line (tensorflo…

…w#361)

dual_net.py

@@ -239,7 +239,6 @@ def eval_metrics_host_call_fn(policy_output, value_output, pi_tensor, policy_cos
         # pi_tensor is one_hot when generated from sgfs (for supervised learning)
         # and soft-max when using self-play records. argmax normalizes the two.
         policy_target_top_1 = tf.argmax(pi_tensor, axis=1)
-        policy_output_top_1 = tf.argmax(policy_output, axis=1)
 
         policy_output_in_top1 = tf.to_float(
             tf.nn.in_top_k(policy_output, policy_target_top_1, k=1))
@@ -251,7 +250,6 @@ def eval_metrics_host_call_fn(policy_output, value_output, pi_tensor, policy_cos
             policy_output,
             tf.one_hot(policy_target_top_1, tf.shape(policy_output)[1]))
 
-        # TODO(sethtroisi): For V10 add tf.variable_scope for tf.metrics.mean's
         with tf.variable_scope("metrics"):
             metric_ops = {
                 'policy_cost': tf.metrics.mean(policy_cost),
@@ -485,6 +483,8 @@ def count_examples(tf_record):
 
         total_examples = sum(map(count_examples, tf_records))
         steps = total_examples // FLAGS.train_batch_size
+        if FLAGS.use_tpu:
+            steps //= FLAGS.num_tpu_cores
 
     if FLAGS.use_tpu:
         def input_fn(params):
@@ -494,7 +494,6 @@ def input_fn(params):
                 random_rotation=True)
         # TODO: get hooks working again with TPUestimator.
         hooks = []
-        steps //= FLAGS.num_tpu_cores
     else:
         def input_fn():
             return preprocessing.get_input_tensors(