Merge pull request #1077 from vince62s/v0.6

Bump version 0.6

CHANGELOG.md

@@ -8,6 +8,10 @@
 
 ### Fixes and improvements
 
+## [0.6.0](https://github.com/OpenNMT/OpenNMT-py/tree/0.6.0) (2018-11-28)
+* Many fixes and code improvements
+* New: Ability to load a yml config file. See examples in config folder.
+
 ## [0.5.0](https://github.com/OpenNMT/OpenNMT-py/tree/0.5.0) (2018-10-24)
 * Fixed advance n_best beam in translate_batch_fast
 * Fixed remove valid set vocab from total vocab

README.md

@@ -5,15 +5,16 @@
 
 This is a [Pytorch](https://github.com/pytorch/pytorch)
 port of [OpenNMT](https://github.com/OpenNMT/OpenNMT),
-an open-source (MIT) neural machine translation system. It is designed to be research friendly to try out new ideas in translation, summary, image-to-text, morphology, and many other domains.
+an open-source (MIT) neural machine translation system. It is designed to be research friendly to try out new ideas in translation, summary, image-to-text, morphology, and many other domains. Some companies have proven the code to be production ready.
 
-Codebase is relatively stable, but PyTorch is still evolving. We currently only support PyTorch 0.4.1 and recommend forking if you need to have stable code.
-
-OpenNMT-py is run as a collaborative open-source project. It is maintained by [Sasha Rush](http://github.com/srush) (Cambridge, MA), [Ben Peters](http://github.com/bpopeters) (Lisbon), and [Jianyu Zhan](http://github.com/jianyuzhan) (Shanghai). The original code was written by [Adam Lerer](http://github.com/adamlerer) (NYC). 
 We love contributions. Please consult the Issues page for any [Contributions Welcome](https://github.com/OpenNMT/OpenNMT-py/issues?q=is%3Aissue+is%3Aopen+label%3A%22contributions+welcome%22) tagged post. 
 
 <center style="padding: 40px"><img width="70%" src="http://opennmt.github.io/simple-attn.png" /></center>
 
+Before raising an issue, make sure you read the requirements and the documentation examples.
+
+Unless there is a bug, please use the [Forum](http://forum.opennmt.net) or [Gitter](https://gitter.im/OpenNMT/OpenNMT-py) to ask questions.
+
 
 Table of Contents
 =================
@@ -22,6 +23,7 @@ Table of Contents
   * [Features](#features)
   * [Quickstart](#quickstart)
   * [Run on FloydHub](#run-on-floydhub)
+  * [Acknowledgements](#acknowledgements)
   * [Citation](#citation)
 
 ## Requirements
@@ -36,7 +38,7 @@ Note that we currently only support PyTorch 0.4.1
 
 ## Features
 
-The following OpenNMT features are implemented:
+Key features:
 
 - [data preprocessing](http://opennmt.net/OpenNMT-py/options/preprocess.html)
 - [Inference (translation) with batching and beam search](http://opennmt.net/OpenNMT-py/options/translate.html)
@@ -130,13 +132,29 @@ The following pretrained models can be downloaded and used with translate.py.
 
 http://opennmt.net/Models-py/
 
+## Acknowledgements
 
+OpenNMT-py is run as a collaborative open-source project.
+The original code was written by [Adam Lerer](http://github.com/adamlerer) (NYC) to reproduce OpenNMT-Lua using Pytorch.
+
+Major contributors are:
+[Sasha Rush](https://github.com/srush) (Cambridge, MA)
+[Vincent Nguyen]((https://github.com/vince62s) (Ubiqus)
+[Ben Peters](http://github.com/bpopeters) (Lisbon)
+[Sebastian Gehrmann](https://github.com/sebastianGehrmann) (Harvard NLP)
+[Yuntian Deng](https://github.com/da03) (Harvard NLP)
+[Guillaume Klein](https://github.com/guillaumekln) (Systran)
+[Paul Tardy](https://github.com/pltrdy) (Ubiqus / Lium)
+[François Hernandez](https://github.com/francoishernandez) (Ubiqus)
+[Jianyu Zhan](http://github.com/jianyuzhan) (Shanghai)
+and more !
+
+OpentNMT-py belongs to the OpenNMT project along with OpenNMT-Lua and OpenNMT-tf.
 
 ## Citation
 
 [OpenNMT: Neural Machine Translation Toolkit](https://arxiv.org/pdf/1805.11462)
 
-
 [OpenNMT technical report](https://doi.org/10.18653/v1/P17-4012)
 
 ```

docs/source/Summarization.md

@@ -47,7 +47,7 @@ python preprocess.py -train_src data/cnndm/train.txt.src \
                      -tgt_seq_length_trunc 100 \
                      -dynamic_dict \
                      -share_vocab \
-                     -max_shard_size (500 * 1024 * 1024)
+                     -shard_size 100000
 ```
 
 (2) Gigaword
@@ -61,7 +61,7 @@ python preprocess.py -train_src data/giga/train.article.txt \
                      -src_seq_length 10000 \
                      -dynamic_dict \
                      -share_vocab \
-                     -max_shard_size (500 * 1024 * 1024)
+                     -shard_size 100000
 ```
 
 
@@ -80,7 +80,7 @@ The training procedure described in this section for the most part follows param
 
 
 We are using using a 128-dimensional word-embedding, and 512-dimensional 1 layer LSTM. On the encoder side, we use a bidirectional LSTM (`brnn`), which means that the 512 dimensions are split into 256 dimensions per direction.
-We also use OpenNMT's default learning rate decay, which halves the learning rate after every epoch once the validation perplexity increased after an epoch (or after epoch 8).
+
 We additionally set the maximum norm of the gradient to 2, and renormalize if the gradient norm exceeds this value and do not use any dropout.
 
 **commands used**:
@@ -96,7 +96,7 @@ python train.py -save_model models/cnndm \
                 -rnn_size 512 \
                 -layers 1 \
                 -encoder_type brnn \
-                -epochs 20 \
+                -train_steps 200000 \
                 -max_grad_norm 2 \
                 -dropout 0. \
                 -batch_size 16 \
@@ -137,7 +137,7 @@ python -u train.py -data data/cnndm/CNNDM \
                    -batch_type tokens \
                    -normalization tokens \
                    -max_generator_batches 2 \
-                   -epochs 25 \
+                   -train_steps 200000 \
                    -start_checkpoint_at 8 \
                    -accum_count 4 \
                    -share_embeddings \
@@ -156,7 +156,7 @@ python train.py -data data/giga/GIGA \
                 -save_model models/giga \
                 -copy_attn \
                 -reuse_copy_attn \
-                -epochs 20
+                -train_steps 200000
 ```
 
 

onmt/__init__.py

@@ -17,4 +17,4 @@
 __all__ = [onmt.inputters, onmt.encoders, onmt.decoders, onmt.models,
            onmt.utils, onmt.modules, "Trainer"]
 
-__version__ = "0.5.0"
+__version__ = "0.6.0"

setup.py

@@ -4,7 +4,7 @@
 
 setup(name='OpenNMT-py',
       description='A python implementation of OpenNMT',
-      version='0.5.0',
+      version='0.6.0',
 
       packages=['onmt', 'onmt.encoders', 'onmt.modules', 'onmt.tests',
                 'onmt.translate', 'onmt.decoders', 'onmt.inputters',