Update comments for two operators. #7457

lcy-seso · 2018-01-11T08:13:37Z

Codes of reorder_lod_tansor_by_rank_op have been updated but its comments has not been updated yet.
Fix the display format of the shrink_rnn_memory_op and update the comments.

guoshengCS · 2018-01-11T13:00:42Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+X = [Seq0, Seq1, Seq2, Seq3]. The indices in RankTable are [3, 0, 2, 1].
+Out =  [Seq3, Seq0, Seq2, Seq1] with a new LoD information.
+
+If the LoD information of Input(X) is empty, this means Input(X) is not a


not a sequence may not be proper. Something like not sequence data might be better.

guoshengCS · 2018-01-11T13:20:39Z

python/paddle/v2/fluid/layers/control_flow.py

@@ -687,11 +687,10 @@ def topk(input, k):


 def lod_tensor_to_array(x, table):
-    """This function performs the operation that converts an LOD_Tensor to
-       an array.
+    """ Convert a LOD_TENSOR_ARRAY to an TensorArray.


Should it be LOD_TENSOR_ARRAY or LoDTensor here

Done. thank you.

guoshengCS · 2018-01-11T13:38:24Z

python/paddle/v2/fluid/layers/control_flow.py


    Args:
-        x (Variable|list): The array that needs to be converted to a tensor.
+        x (Variable|list): The LoD Tensor Array to be converted to a tensor.


Should converted to a tensor be converted to a LoDTensor here

guoshengCS · 2018-01-11T13:41:04Z

python/paddle/v2/fluid/layers/control_flow.py

    Returns:
-        Variable: The tensor type variable that has the data written to it.
+        Variable: The ouput LOD_TENSOR_ARRAY where the input tensor is written.


The ouput should be The output.

abhinavarora · 2018-01-11T18:15:57Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc


-The Out =  [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
+If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the


is -> are

abhinavarora · 2018-01-11T18:16:20Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc


-The Out =  [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
+If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the
+Input(X) will be reordered that the forth sequence in Input(X) will become the


forth -> fourth

abhinavarora

There is some language polishing needed. Apart from that everything looks good. Thank you for the PR!

abhinavarora · 2018-01-11T18:16:40Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

-The Out =  [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
+If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the
+Input(X) will be reordered that the forth sequence in Input(X) will become the
+first one, and then followed by the originally first, third, and the second one.


originally -> original

abhinavarora · 2018-01-11T18:17:34Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+Out =  [Seq3, Seq0, Seq2, Seq1] with a new LoD information.
+
+If the LoD information of Input(X) is empty, this means Input(X) is not a
+sequcence. This is also identical to a batch of sequences each sequence in which


The sentence should be -> This is also identical to a batch of sequences where each sequence has a fixed length 1.

abhinavarora · 2018-01-11T18:17:58Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+indices in RankTable are [3, 0, 2, 1].
+Out = [Slice3, Slice0, Slice2, Slice1] with no LoD information is appended.
+
+NOTE: This operator sorts Input(X) according to a given LoDRankTable which dose


dose -> does

abhinavarora · 2018-01-11T18:19:24Z

python/paddle/v2/fluid/layers/control_flow.py


    Args:
-        x (Variable|list): The tensor that needs to be converted to an array.
+        x (Variable|list): The lod tensor to be converted to a lod tensor array.


lod -> LOD

abhinavarora · 2018-01-11T18:20:11Z

python/paddle/v2/fluid/layers/tensor.py

@@ -146,25 +146,27 @@ def fill_constant(shape, dtype, value, out=None):
    """
    **fill_constant**

-    This function creates a tensor of specified *shape* and
-    *dtype*, and initializes this with a constant supplied in *value*.
+    This function creates a tensor with the specified `shape` and


You can drop the

abhinavarora · 2018-01-11T18:20:30Z

python/paddle/v2/fluid/layers/tensor.py

-    This function creates a tensor of specified *shape* and
-    *dtype*, and initializes this with a constant supplied in *value*.
+    This function creates a tensor with the specified `shape` and
+    `dtype`, and initializes it with the constant specifed by `value`.


lcy-seso

follow comments. Thanks for all the corrections.

lcy-seso · 2018-01-12T03:41:32Z

python/paddle/v2/fluid/layers/tensor.py

@@ -146,25 +146,27 @@ def fill_constant(shape, dtype, value, out=None):
    """
    **fill_constant**

-    This function creates a tensor of specified *shape* and
-    *dtype*, and initializes this with a constant supplied in *value*.
+    This function creates a tensor with the specified `shape` and


lcy-seso · 2018-01-12T03:41:40Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc


-The Out =  [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
+If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the


lcy-seso · 2018-01-12T03:41:45Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

-The Out =  [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
+If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the
+Input(X) will be reordered that the forth sequence in Input(X) will become the
+first one, and then followed by the originally first, third, and the second one.


lcy-seso · 2018-01-12T03:41:51Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+X = [Seq0, Seq1, Seq2, Seq3]. The indices in RankTable are [3, 0, 2, 1].
+Out =  [Seq3, Seq0, Seq2, Seq1] with a new LoD information.
+
+If the LoD information of Input(X) is empty, this means Input(X) is not a


lcy-seso · 2018-01-12T03:41:58Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+Out =  [Seq3, Seq0, Seq2, Seq1] with a new LoD information.
+
+If the LoD information of Input(X) is empty, this means Input(X) is not a
+sequcence. This is also identical to a batch of sequences each sequence in which


lcy-seso · 2018-01-12T03:42:03Z

paddle/operators/reorder_lod_tensor_by_rank_op.cc

+indices in RankTable are [3, 0, 2, 1].
+Out = [Slice3, Slice0, Slice2, Slice1] with no LoD information is appended.
+
+NOTE: This operator sorts Input(X) according to a given LoDRankTable which dose


lcy-seso · 2018-01-12T03:42:12Z

python/paddle/v2/fluid/layers/control_flow.py

@@ -687,11 +687,10 @@ def topk(input, k):


 def lod_tensor_to_array(x, table):
-    """This function performs the operation that converts an LOD_Tensor to
-       an array.
+    """ Convert a LOD_TENSOR_ARRAY to an TensorArray.


Done. thank you.

lcy-seso · 2018-01-12T03:42:17Z

python/paddle/v2/fluid/layers/control_flow.py


    Args:
-        x (Variable|list): The tensor that needs to be converted to an array.
+        x (Variable|list): The lod tensor to be converted to a lod tensor array.


lcy-seso · 2018-01-12T03:42:25Z

python/paddle/v2/fluid/layers/control_flow.py

    Returns:
-        Variable: The tensor type variable that has the data written to it.
+        Variable: The ouput LOD_TENSOR_ARRAY where the input tensor is written.


lcy-seso · 2018-01-12T03:42:32Z

python/paddle/v2/fluid/layers/tensor.py

-    This function creates a tensor of specified *shape* and
-    *dtype*, and initializes this with a constant supplied in *value*.
+    This function creates a tensor with the specified `shape` and
+    `dtype`, and initializes it with the constant specifed by `value`.


lcy-seso added 5 commits January 3, 2018 09:16

update code comments.

67d508d

Merge branch 'develop' into update_comments

2aca8c7

Merge branch 'develop' into update_comments

7829031

Merge branch 'develop' into update_comments

5b912e2

Merge branch 'develop' into update_comments

e3e9a40

lcy-seso requested review from abhinavarora, guoshengCS and pkuyym January 11, 2018 08:13

update the comments.

49c20a0

lcy-seso force-pushed the update_comments branch 2 times, most recently from a2be67c to c381560 Compare January 11, 2018 13:10

Merge branch 'develop' into update_comments

473fd63

lcy-seso force-pushed the update_comments branch from c381560 to 473fd63 Compare January 11, 2018 13:13

guoshengCS reviewed Jan 11, 2018

View reviewed changes

abhinavarora reviewed Jan 11, 2018

View reviewed changes

abhinavarora suggested changes Jan 11, 2018

View reviewed changes

follow comments.

40ed8a0

lcy-seso commented Jan 12, 2018

View reviewed changes

abhinavarora approved these changes Jan 12, 2018

View reviewed changes

abhinavarora merged commit db65f49 into PaddlePaddle:develop Jan 12, 2018

lcy-seso deleted the update_comments branch January 16, 2018 11:52

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update comments for two operators. #7457

Update comments for two operators. #7457

lcy-seso commented Jan 11, 2018 •

edited

Loading

guoshengCS Jan 11, 2018

lcy-seso Jan 12, 2018

guoshengCS Jan 11, 2018

lcy-seso Jan 12, 2018

guoshengCS Jan 11, 2018

guoshengCS Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

abhinavarora left a comment

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

abhinavarora Jan 11, 2018

lcy-seso Jan 12, 2018

lcy-seso left a comment

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018

lcy-seso Jan 12, 2018


		The Out = [Seq3, Seq0, Seq2, Seq1] with correct LoD information.
		If the indices stored in the Input(RankTable) is [3, 0, 2, 1], the

Update comments for two operators. #7457

Update comments for two operators. #7457

Conversation

lcy-seso commented Jan 11, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

abhinavarora left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcy-seso left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcy-seso commented Jan 11, 2018 •

edited

Loading