Remove set_shape as tfhub module updated.

PiperOrigin-RevId: 290977517

Remove set_shape as tfhub module updated.
PiperOrigin-RevId: 290977517
5ed215b2 · Hongkun Yu · A. Unique TensorFlower · d6117ed1 · 5ed215b2
Commit 5ed215b2 authored Jan 22, 2020 by Hongkun Yu Committed by A. Unique TensorFlower Jan 22, 2020
Hide whitespace changes
Inline Side-by-side

Showing with 0 additions and 3 deletions

official/nlp/bert_models.py official/nlp/bert_models.py +0 -3

No files found.
--- a/official/nlp/bert_models.py
+++ b/official/nlp/bert_models.py
@@ -309,9 +309,6 @@ def squad_model(bert_config,
  core_model = hub.KerasLayer(hub_module_url, trainable=True)
  _, sequence_output = core_model(
      [input_word_ids, input_mask, input_type_ids])
-  # Sets the shape manually due to a bug in TF shape inference.
-  # TODO(hongkuny): remove this once shape inference is correct.
-  sequence_output.set_shape((None, max_seq_length, bert_config.hidden_size))

  squad_logits_layer = BertSquadLogitsLayer(
      initializer=initializer, float_type=float_type, name='squad_logits')