Update to MXNet version 0.12.1. (#213)

awslabs · Nov 20, 2017 · c3c274a · c3c274a
1 parent 556631d
commit c3c274a
Show file tree

Hide file tree

Showing 8 changed files with 12 additions and 7 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,6 +10,11 @@ Note that Sockeye has checks in place to not translate with an old model that wa
 
 For each item we will potentially have subsections for: _Added_, _Changed_, _Removed_, _Deprecated_, and _Fixed_.
 
+## [1.12.2]
+### Changed
+ - Updated to [MXNet 0.12.1](https://github.com/apache/incubator-mxnet/releases/tag/0.12.1), which includes an important
+ bug fix for CPU decoding.
+
 ## [1.12.1]
 ### Changed
  - Removed dependency on sacrebleu pip package. Now imports directly from `contrib/`.

diff --git a/README.md b/README.md
@@ -18,7 +18,7 @@ Recent developments and changes are tracked in our [changelog](https://github.co
 
 Sockeye requires:
 - **Python3**
-- [MXNet-0.12.0](https://github.com/dmlc/mxnet/tree/v0.12.0)
+- [MXNet-0.12.1](https://github.com/apache/incubator-mxnet/tree/0.12.1)
 - numpy
 
 ## Installation

diff --git a/requirements.gpu-cu75.txt b/requirements.gpu-cu75.txt
@@ -1,4 +1,4 @@
 pyyaml
-mxnet-cu75==0.12.0
+mxnet-cu75==0.12.1
 numpy>=1.12
 typing
diff --git a/requirements.gpu-cu80.txt b/requirements.gpu-cu80.txt
@@ -1,4 +1,4 @@
 pyyaml
-mxnet-cu80==0.12.0
+mxnet-cu80==0.12.1
 numpy>=1.12
 typing
diff --git a/requirements.gpu-cu90.txt b/requirements.gpu-cu90.txt
@@ -1,4 +1,4 @@
 pyyaml
-mxnet-cu90==0.12.0
+mxnet-cu90==0.12.1
 numpy>=1.12
 typing
diff --git a/requirements.txt b/requirements.txt
@@ -1,4 +1,4 @@
 pyyaml
-mxnet==0.12.0
+mxnet==0.12.1
 numpy>=1.12
 typing
diff --git a/sockeye/__init__.py b/sockeye/__init__.py
@@ -11,4 +11,4 @@
 # express or implied. See the License for the specific language governing
 # permissions and limitations under the License.
 
-__version__ = '1.12.1'
+__version__ = '1.12.2'
diff --git a/test/system/test_seq_copy_sys.py b/test/system/test_seq_copy_sys.py
@@ -54,7 +54,7 @@
     ("Copy:transformer:lstm",
      "--encoder transformer --num-layers 2:1 --rnn-cell-type lstm --rnn-num-hidden 64 --num-embed 32"
      " --rnn-attention-type mhdot --rnn-attention-num-hidden 32 --batch-size 16 --rnn-attention-mhdot-heads 1"
-     " --loss cross-entropy --optimized-metric perplexity --max-updates 5000"
+     " --loss cross-entropy --optimized-metric perplexity --max-updates 6000"
      " --transformer-attention-heads 4 --transformer-model-size 64"
      " --transformer-feed-forward-num-hidden 64"
      " --checkpoint-frequency 1000 --optimizer adam --initial-learning-rate 0.001",