Fix some bugs of dygraph Transformer.

dygraph/transformer/model.py

@@ -21,8 +21,6 @@ import paddle.fluid.layers as layers
 from paddle.fluid.dygraph import Embedding, LayerNorm, Linear, Layer, to_variable
 from paddle.fluid.dygraph.learning_rate_scheduler import LearningRateDecay
 
-from config import word_emb_param_names, pos_enc_param_names
-
 
 def position_encoding_init(n_position, d_pos_vec):
     """
@@ -138,7 +136,7 @@ class MultiHeadAttention(Layer):
         q = layers.reshape(x=q, shape=[0, 0, self.n_head, self.d_key])
         q = layers.transpose(x=q, perm=[0, 2, 1, 3])
 
-        if cache is not None and static_kv and cache.has_key("static_k"):
+        if cache is not None and static_kv and "static_k" in cache:
             # for encoder-decoder attention in inference and has cached
             k = cache["static_k"]
             v = cache["static_v"]
@@ -151,7 +149,7 @@ class MultiHeadAttention(Layer):
             v = layers.transpose(x=v, perm=[0, 2, 1, 3])
 
         if cache is not None:
-            if static_kv and not cache.has_key("static_k"):
+            if static_kv and not "static_k" in cache:
                 # for encoder-decoder attention in inference and has not cached
                 cache["static_k"], cache["static_v"] = k, v
             elif not static_kv:
@@ -180,7 +178,7 @@ class MultiHeadAttention(Layer):
                                      dropout_prob=self.dropout_rate,
                                      is_test=False)
 
-            out = layers.matmul(weights, v)
+        out = layers.matmul(weights, v)
 
         # combine heads
         out = layers.transpose(out, perm=[0, 2, 1, 3])
@@ -817,15 +815,15 @@ class Transformer(Layer):
         return finished_seq, finished_scores
 
     def beam_search(self,
-                     src_word,
-                     src_pos,
-                     src_slf_attn_bias,
-                     trg_word,
-                     trg_src_attn_bias,
-                     bos_id=0,
-                     eos_id=1,
-                     beam_size=4,
-                     max_len=256):
+                    src_word,
+                    src_pos,
+                    src_slf_attn_bias,
+                    trg_word,
+                    trg_src_attn_bias,
+                    bos_id=0,
+                    eos_id=1,
+                    beam_size=4,
+                    max_len=256):
         if beam_size == 1:
             return self._greedy_search(src_word,
                                        src_pos,
@@ -1017,6 +1015,9 @@ class Transformer(Layer):
         trg_word = layers.fill_constant(shape=[batch_size, 1],
                                         dtype="int64",
                                         value=bos_id)
+        finished = layers.fill_constant(shape=[batch_size, 1],
+                                        dtype="bool",
+                                        value=0)
 
         ## init states (caches) for transformer
         caches = [{
@@ -1045,7 +1046,8 @@ class Transformer(Layer):
             scores = log_probs
             topk_scores, topk_indices = layers.topk(input=scores, k=1)
 
-            finished = layers.equal(topk_indices, end_token_tensor)
+            finished = layers.logical_or(
+                finished, layers.equal(topk_indices, end_token_tensor))
             trg_word = topk_indices
             log_probs = topk_scores
 

dygraph/transformer/train.py

@@ -145,7 +145,6 @@ def do_train(args):
         # train loop
         for pass_id in range(args.epoch):
             pass_start_time = time.time()
-            avg_batch_time = time.time()
             batch_id = 0
             for input_data in train_loader():
                 (src_word, src_pos, src_slf_attn_bias, trg_word, trg_pos,
@@ -178,6 +177,7 @@ def do_train(args):
                             (step_idx, pass_id, batch_id, total_avg_cost,
                             total_avg_cost - loss_normalizer,
                             np.exp([min(total_avg_cost, 100)])))
+                        avg_batch_time = time.time()
                     else:
                         logging.info(
                             "step_idx: %d, epoch: %d, batch: %d, avg loss: %f, "
@@ -186,6 +186,7 @@ def do_train(args):
                             total_avg_cost - loss_normalizer,
                             np.exp([min(total_avg_cost, 100)]),
                             args.print_step / (time.time() - avg_batch_time)))
+                        avg_batch_time = time.time()
 
 
                 if step_idx % args.save_step == 0 and step_idx != 0:
@@ -228,7 +229,6 @@ def do_train(args):
                             optimizer.state_dict(),
                             os.path.join(model_dir, "transformer"))
 
-                avg_batch_time = time.time()
                 batch_id += 1
                 step_idx += 1