1. 10 11月, 2017 3 次提交
    • Y
      feature/while_op (#5502) · 40367d18
      Yang Yang(Tony) 提交于
      * first commit
      
      * Python API for while op
      
      * Python Unittest for simple while_op forward
      
      * fix out to be list
      
      * Fix UT
      
      * VarType
      
      * Fix several bugs
      
      * Fix bug
      
      * Fix bug
      
      * Fix Bug
      
      * Fix bug
      
      * Fix unittest
      
      * Remove debug log
      
      * Add comments
      
      * add PADDLE_ENFORCE
      
      * while_grad_op first commit
      
      * Add `BlockDescBind::FindRecursiveOrCreateVar()` and fix bugs
      
      * refine code
      
      * fix unittest bug
      40367d18
    • K
      Adding documentation for every function in layers.py (#5529) · b5901a3a
      kavyasrinet 提交于
      * Adding operator assignment
      
      * Adding documentation to layers.py
      
      * Removing file from another PR
      b5901a3a
    • S
      Fix attribute naming for momentum_op (#5453) · 2e355f03
      Siddharth Goyal 提交于
      * Fix attribute naming for momentum_op
      
      * Fix minor typo in comment
      
      * Fix attribute name
      
      * Fix names in test_optimizer
      
      * Fix python wrapper
      2e355f03
  2. 09 11月, 2017 7 次提交
    • D
      follow comments. · d60fe75a
      dangqingqing 提交于
      d60fe75a
    • Y
      Do not sum output if that output is not a gradient · c9fc7ba9
      Yang Yu 提交于
      * increament is default inplace
      c9fc7ba9
    • H
      Remove unused g_main_program in tests · b8a20432
      Helin Wang 提交于
      b8a20432
    • F
      Add grad for lodtensor array ops (#5461) · b698d19b
      fengjiayi 提交于
      * Add LoDRankTable
      
      LoD Rank Table stores the `level` of `lod` which is ordered by sequence
      length in descending order. It is useful when implement dynamic RNN and
      is shared by dynamic RNN memory, dynamic RNN slice input and dynamic
      RNN slice output operators.
      
      * Add skeleton for array_to_lod_tensor and lod_tensor_to_array
      
      * Add VarType::LoDTensorArray
      
      * Add PyBind of LoDTensorArray
      
      * Add InferVarType
      
      * Add first unittest
      
      * Add ut
      
      * Add unittest
      
      * Add unittest
      
      * Add unittests
      
      * update
      
      * init
      
      * add infershape for lod_tensor_to_array_op
      
      * compelete array_to_lod_tensor_op
      
      * copy data
      
      * clean code
      
      * clean code
      
      * Fix unittest data
      
      * fix bugs
      
      * fix compile error
      
      * Refine TensorToArrayOp
      
      * refactor array_to_lod_tensor
      
      * Unittest
      
      * fix bugs
      
      * Fix unittest
      
      * Fix unittest
      
      * debug
      
      * Debug
      
      * Fix unittest
      
      * Add grad for ops
      
      * Debug
      
      * Fix a bug
      
      * fix a bug
      
      * fix a bug
      b698d19b
    • Y
      Add `lod_array_length` operator · d24d8c20
      Yang Yu 提交于
      d24d8c20
    • Y
      Add increment op · 6d41bfb7
      Yang Yu 提交于
      6d41bfb7
    • G
      c8dcd9a9
  3. 08 11月, 2017 11 次提交
    • L
      add doc for image.py · 15133229
      Luo Tao 提交于
      15133229
    • Y
      Static lstm sanity check (#5365) · 870650d8
      Yang Yang(Tony) 提交于
      * add fill_constant_batch_size_like_op to rnn h_boot
      
      * first commit
      
      * merge develop; fix conflict
      
      * update to main_program
      870650d8
    • T
      fix accuracy cudamemset · 6308ccc2
      typhoonzero 提交于
      6308ccc2
    • Y
      Rename shrink_state -> shrink_rnn_memory · 01425309
      Yang Yu 提交于
      Follow comments
      01425309
    • Y
      Fix Unittest · b4dddb29
      Yang Yu 提交于
      b4dddb29
    • W
      ac7cca18
    • C
      fix attr name · cdf5e871
      chengduoZH 提交于
      cdf5e871
    • Y
      Feature/rnn to array to lod tensor (#5411) · f72729d4
      Yu Yang 提交于
      * Add LoDRankTable
      
      LoD Rank Table stores the `level` of `lod` which is ordered by sequence
      length in descending order. It is useful when implement dynamic RNN and
      is shared by dynamic RNN memory, dynamic RNN slice input and dynamic
      RNN slice output operators.
      
      * Add skeleton for array_to_lod_tensor and lod_tensor_to_array
      
      * Add VarType::LoDTensorArray
      
      * Add PyBind of LoDTensorArray
      
      * Add InferVarType
      
      * Add first unittest
      
      * Add ut
      
      * Add unittest
      
      * Add unittest
      
      * Add unittests
      
      * update
      
      * init
      
      * add infershape for lod_tensor_to_array_op
      
      * compelete array_to_lod_tensor_op
      
      * copy data
      
      * clean code
      
      * clean code
      
      * Fix unittest data
      
      * fix bugs
      
      * fix compile error
      
      * Refine TensorToArrayOp
      
      * refactor array_to_lod_tensor
      
      * Unittest
      
      * fix bugs
      
      * Fix unittest
      
      * Fix unittest
      
      * debug
      
      * Debug
      
      * Fix unittest
      
      * clean code
      
      * refactor
      
      * use ostream
      
      * update test
      
      * fix gpu build error
      
      * make gpu test pass
      f72729d4
    • Y
      Shrink State Operator · 2dd91dd5
      Yang Yu 提交于
      Used for shrink memories state in DyRNN. The height of state could
      be shrinked after running a step block.
      2dd91dd5
    • Y
      Temporary disable accurary_op test (#5451) · d9e5eba0
      Yu Yang 提交于
      d9e5eba0
    • Y
      Compare Operator (#5325) · f74fb790
      Yu Yang 提交于
      * Compare Operator
      
      * Follow comments
      f74fb790
  4. 07 11月, 2017 5 次提交
    • T
      update · fc4d4b88
      typhoonzero 提交于
      fc4d4b88
    • D
      Enhance unit testing. · d94c936b
      dangqingqing 提交于
      1. user can disable peephole connections.
      2. not calculate some gradients.
      d94c936b
    • Y
      Add unittest, backward of array read/write op (#5409) · 6cde889b
      Yu Yang 提交于
      * Use stable_sort in lod_rank_table
      
      It is easy to debug and test when use `stable_sort`and the time
      complexity is not changed.
      
      * Add LoDTensorArray
      
      * Stash
      
      * Better debug message for IsInitialized
      
      * Stash
      
      * Better debug message for IsInitialized
      
      * Complete array read/write op unittests
      
      * Add unittest, Gradient of array read/write
      
      * Follow comments
      6cde889b
    • Y
    • Y
      ReadFromArray/WriteToArray op (#5407) · c9b57dcc
      Yu Yang 提交于
      * Use stable_sort in lod_rank_table
      
      It is easy to debug and test when use `stable_sort`and the time
      complexity is not changed.
      
      * Add LoDTensorArray
      
      * Stash
      
      * Better debug message for IsInitialized
      
      * Stash
      
      * Better debug message for IsInitialized
      
      * Complete array read/write op unittests
      c9b57dcc
  5. 06 11月, 2017 8 次提交
  6. 05 11月, 2017 2 次提交
  7. 04 11月, 2017 3 次提交
    • C
      Add the crf_decoding operator. (#5352) · 45eabb8c
      Cao Ying 提交于
      * proj init.
      
      * add unittest and implementation.
      45eabb8c
    • Q
      Add acc test to image classification (#5336) · 906e2565
      Qiao Longfei 提交于
      * add acc layer
      * memory log level change from 3 to 10
      * use gaussian random to init conv parameters
      * use initializer
      * fix import
      * batch_norm use helper to create persistable var
      * refine code
      * train only 2 batches for test
      * use g_program and g_init_program
      * use XavierInitializer to init fc parameter
      906e2565
    • Y
      Add LoDRankTable (#5349) · 74849158
      Yu Yang 提交于
      * Add LoDRankTable
      
      LoD Rank Table stores the `level` of `lod` which is ordered by sequence
      length in descending order. It is useful when implement dynamic RNN and
      is shared by dynamic RNN memory, dynamic RNN slice input and dynamic
      RNN slice output operators.
      
      * Add InferVarType
      74849158
  8. 03 11月, 2017 1 次提交