[](https://github.com/OpenNMT/OpenNMT-tf/actions?query=workflow%3ACI) [](https://codecov.io/gh/OpenNMT/OpenNMT-tf) [](https://badge.fury.io/py/OpenNMT-tf) [](https://opennmt.net/OpenNMT-tf/) [](https://gitter.im/OpenNMT/OpenNMT-tf?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge) [](https://forum.opennmt.net/)
# OpenNMT-tf
OpenNMT-tf is a general purpose sequence learning toolkit using TensorFlow 2. While neural machine translation is the main target task, it has been designed to more generally support:
* sequence to sequence mapping
* sequence tagging
* sequence classification
* language modeling
The project is production-oriented and comes with [backward compatibility guarantees](https://github.com/OpenNMT/OpenNMT-tf/blob/master/CHANGELOG.md).
## Key features
### Modular model architecture
Models are described with code to allow training custom architectures and overriding default behavior. For example, the following instance defines a sequence to sequence model with 2 concatenated input features, a self-attentional encoder, and an attentional RNN decoder sharing its input and output embeddings:
```python
opennmt.models.SequenceToSequence(
source_inputter=opennmt.inputters.ParallelInputter(
[
opennmt.inputters.WordEmbedder(embedding_size=256),
opennmt.inputters.WordEmbedder(embedding_size=256),
],
reducer=opennmt.layers.ConcatReducer(axis=-1),
),
target_inputter=opennmt.inputters.WordEmbedder(embedding_size=512),
encoder=opennmt.encoders.SelfAttentionEncoder(num_layers=6),
decoder=opennmt.decoders.AttentionalRNNDecoder(
num_layers=4,
num_units=512,
attention_mechanism_class=tfa.seq2seq.LuongAttention,
),
share_embeddings=opennmt.models.EmbeddingsSharingLevel.TARGET,
)
```
The [`opennmt`](https://opennmt.net/OpenNMT-tf/package/overview.html) package exposes other building blocks that can be used to design:
* [multiple input features](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.ParallelInputter.html)
* [mixed embedding representation](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.MixedInputter.html)
* [multi-source context](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.ParallelInputter.html)
* [cascaded](https://opennmt.net/OpenNMT-tf/package/opennmt.encoders.SequentialEncoder.html) or [multi-column](https://opennmt.net/OpenNMT-tf/package/opennmt.encoders.ParallelEncoder.html) encoder
* [hybrid sequence to sequence models](https://opennmt.net/OpenNMT-tf/package/opennmt.models.SequenceToSequence.html)
Standard models such as the Transformer are defined in a [model catalog](https://github.com/OpenNMT/OpenNMT-tf/blob/master/opennmt/models/catalog.py) and can be used without additional configuration.
*Find more information about model configuration in the [documentation](https://opennmt.net/OpenNMT-tf/model.html).*
### Full TensorFlow 2 integration
OpenNMT-tf is fully integrated in the TensorFlow 2 ecosystem:
* Reusable layers extending [`tf.keras.layers.Layer`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Layer)
* Multi-GPU training with [`tf.distribute`](https://www.tensorflow.org/api_docs/python/tf/distribute) and distributed training with [Horovod](https://github.com/horovod/horovod)
* Mixed precision training with [`tf.keras.mixed_precision`](https://www.tensorflow.org/guide/mixed_precision)
* Visualization with [TensorBoard](https://www.tensorflow.org/tensorboard)
* `tf.function` graph tracing that can be [exported to a SavedModel](https://opennmt.net/OpenNMT-tf/serving.html) and served with [TensorFlow Serving](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/serving/tensorflow_serving) or [Python](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/serving/python)
### Compatibility with CTranslate2
[CTranslate2](https://github.com/OpenNMT/CTranslate2) is an optimized inference engine for OpenNMT models featuring fast CPU and GPU execution, model quantization, parallel translations, dynamic memory usage, interactive decoding, and more! OpenNMT-tf can [automatically export](https://opennmt.net/OpenNMT-tf/serving.html#ctranslate2) models to be used in CTranslate2.
### Dynamic data pipeline
OpenNMT-tf does not require to compile the data before the training. Instead, it can directly read text files and preprocess the data when needed by the training. This allows [on-the-fly tokenization](https://opennmt.net/OpenNMT-tf/tokenization.html) and data augmentation by injecting random noise.
### Model fine-tuning
OpenNMT-tf supports model fine-tuning workflows:
* Model weights can be transferred to new word vocabularies, e.g. to inject domain terminology before fine-tuning on in-domain data
* [Contrastive learning](https://ai.google/research/pubs/pub48253/) to reduce word omission errors
### Source-target alignment
Sequence to sequence models can be trained with [guided alignment](https://arxiv.org/abs/1607.01628) and alignment information are returned as part of the translation API.
---
OpenNMT-tf also implements most of the techniques commonly used to train and evaluate sequence models, such as:
* automatic evaluation during the training
* multiple decoding strategy: greedy search, beam search, random sampling
* N-best rescoring
* gradient accumulation
* scheduled sampling
* checkpoint averaging
* ... and more!
*See the [documentation](https://opennmt.net/OpenNMT-tf/) to learn how to use these features.*
## Usage
OpenNMT-tf requires:
* Python 3.7 or above
* TensorFlow 2.6, 2.7, 2.8, 2.9, 2.10, 2.11, 2.12, or 2.13
We recommend installing it with `pip`:
```bash
pip install --upgrade pip
pip install OpenNMT-tf
```
*See the [documentation](https://opennmt.net/OpenNMT-tf/installation.html) for more information.*
### Command line
OpenNMT-tf comes with several command line utilities to prepare data, train, and evaluate models.
For all tasks involving a model execution, OpenNMT-tf uses a unique entrypoint: `onmt-main`. A typical OpenNMT-tf run consists of 3 elements:
* the **model** type
* the **parameters** described in a YAML file
* the **run** type such as `train`, `eval`, `infer`, `export`, `score`, `average_checkpoints`, or `update_vocab`
that are passed to the main script:
```
onmt-main --model_type <model> --config <config_file.yml> --auto_config <run_type> <run_options>
```
*For more information and examples on how to use OpenNMT-tf, please visit [our documentation](https://opennmt.net/OpenNMT-tf).*
### Library
OpenNMT-tf also exposes [well-defined and stable APIs](https://opennmt.net/OpenNMT-tf/package/overview.html), from high-level training utilities to low-level model layers and dataset transformations.
For example, the `Runner` class can be used to train and evaluate models with few lines of code:
```python
import opennmt
config = {
"model_dir": "/data/wmt-ende/checkpoints/",
"data": {
"source_vocabulary": "/data/wmt-ende/joint-vocab.txt",
"target_vocabulary": "/data/wmt-ende/joint-vocab.txt",
"train_features_file": "/data/wmt-ende/train.en",
"train_labels_file": "/data/wmt-ende/train.de",
"eval_features_file": "/data/wmt-ende/valid.en",
"eval_labels_file": "/data/wmt-ende/valid.de",
}
}
model = opennmt.models.TransformerBase()
runner = opennmt.Runner(model, config, auto_config=True)
runner.train(num_devices=2, with_eval=True)
```
Here is another example using OpenNMT-tf to run efficient beam search with a self-attentional decoder:
```python
decoder = opennmt.decoders.SelfAttentionDecoder(num_layers=6, vocab_size=32000)
initial_state = decoder.initial_state(
memory=memory, memory_sequence_length=memory_sequence_length
)
batch_size = tf.shape(memory)[0]
start_ids = tf.fill([batch_size], opennmt.START_OF_SENTENCE_ID)
decoding_result = decoder.dynamic_decode(
target_embedding,
start_ids=start_ids,
initial_state=initial_state,
decoding_strategy=opennmt.utils.BeamSearch(4),
)
```
More examples using OpenNMT-tf as a library can be found online:
* The directory [examples/library](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/library) contains additional examples that use OpenNMT-tf as a library
* [nmt-wizard-docker](https://github.com/OpenNMT/nmt-wizard-docker) uses the high-level `opennmt.Runner` API to wrap OpenNMT-tf with a custom interface for training, translating, and serving
*For a complete overview of the APIs, see the [package documentation](https://opennmt.net/OpenNMT-tf/package/overview.html).*
## Additional resources
* [Documentation](https://opennmt.net/OpenNMT-tf)
* [Forum](https://forum.opennmt.net)
* [Gitter](https://gitter.im/OpenNMT/OpenNMT-tf)
Raw data
{
"_id": null,
"home_page": "https://opennmt.net",
"name": "OpenNMT-tf",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.7",
"maintainer_email": "",
"keywords": "tensorflow opennmt nmt neural machine translation",
"author": "OpenNMT",
"author_email": "",
"download_url": "https://files.pythonhosted.org/packages/39/a7/aa003550a746f2c843718b55f21a6697a61b2c9e10587c5ba0fa4b62708b/OpenNMT-tf-2.32.0.tar.gz",
"platform": null,
"description": "[](https://github.com/OpenNMT/OpenNMT-tf/actions?query=workflow%3ACI) [](https://codecov.io/gh/OpenNMT/OpenNMT-tf) [](https://badge.fury.io/py/OpenNMT-tf) [](https://opennmt.net/OpenNMT-tf/) [](https://gitter.im/OpenNMT/OpenNMT-tf?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge) [](https://forum.opennmt.net/)\n\n# OpenNMT-tf\n\nOpenNMT-tf is a general purpose sequence learning toolkit using TensorFlow 2. While neural machine translation is the main target task, it has been designed to more generally support:\n\n* sequence to sequence mapping\n* sequence tagging\n* sequence classification\n* language modeling\n\nThe project is production-oriented and comes with [backward compatibility guarantees](https://github.com/OpenNMT/OpenNMT-tf/blob/master/CHANGELOG.md).\n\n## Key features\n\n### Modular model architecture\n\nModels are described with code to allow training custom architectures and overriding default behavior. For example, the following instance defines a sequence to sequence model with 2 concatenated input features, a self-attentional encoder, and an attentional RNN decoder sharing its input and output embeddings:\n\n```python\nopennmt.models.SequenceToSequence(\n source_inputter=opennmt.inputters.ParallelInputter(\n [\n opennmt.inputters.WordEmbedder(embedding_size=256),\n opennmt.inputters.WordEmbedder(embedding_size=256),\n ],\n reducer=opennmt.layers.ConcatReducer(axis=-1),\n ),\n target_inputter=opennmt.inputters.WordEmbedder(embedding_size=512),\n encoder=opennmt.encoders.SelfAttentionEncoder(num_layers=6),\n decoder=opennmt.decoders.AttentionalRNNDecoder(\n num_layers=4,\n num_units=512,\n attention_mechanism_class=tfa.seq2seq.LuongAttention,\n ),\n share_embeddings=opennmt.models.EmbeddingsSharingLevel.TARGET,\n)\n```\n\nThe [`opennmt`](https://opennmt.net/OpenNMT-tf/package/overview.html) package exposes other building blocks that can be used to design:\n\n* [multiple input features](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.ParallelInputter.html)\n* [mixed embedding representation](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.MixedInputter.html)\n* [multi-source context](https://opennmt.net/OpenNMT-tf/package/opennmt.inputters.ParallelInputter.html)\n* [cascaded](https://opennmt.net/OpenNMT-tf/package/opennmt.encoders.SequentialEncoder.html) or [multi-column](https://opennmt.net/OpenNMT-tf/package/opennmt.encoders.ParallelEncoder.html) encoder\n* [hybrid sequence to sequence models](https://opennmt.net/OpenNMT-tf/package/opennmt.models.SequenceToSequence.html)\n\nStandard models such as the Transformer are defined in a [model catalog](https://github.com/OpenNMT/OpenNMT-tf/blob/master/opennmt/models/catalog.py) and can be used without additional configuration.\n\n*Find more information about model configuration in the [documentation](https://opennmt.net/OpenNMT-tf/model.html).*\n\n### Full TensorFlow 2 integration\n\nOpenNMT-tf is fully integrated in the TensorFlow 2 ecosystem:\n\n* Reusable layers extending [`tf.keras.layers.Layer`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Layer)\n* Multi-GPU training with [`tf.distribute`](https://www.tensorflow.org/api_docs/python/tf/distribute) and distributed training with [Horovod](https://github.com/horovod/horovod)\n* Mixed precision training with [`tf.keras.mixed_precision`](https://www.tensorflow.org/guide/mixed_precision)\n* Visualization with [TensorBoard](https://www.tensorflow.org/tensorboard)\n* `tf.function` graph tracing that can be [exported to a SavedModel](https://opennmt.net/OpenNMT-tf/serving.html) and served with [TensorFlow Serving](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/serving/tensorflow_serving) or [Python](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/serving/python)\n\n### Compatibility with CTranslate2\n\n[CTranslate2](https://github.com/OpenNMT/CTranslate2) is an optimized inference engine for OpenNMT models featuring fast CPU and GPU execution, model quantization, parallel translations, dynamic memory usage, interactive decoding, and more! OpenNMT-tf can [automatically export](https://opennmt.net/OpenNMT-tf/serving.html#ctranslate2) models to be used in CTranslate2.\n\n### Dynamic data pipeline\n\nOpenNMT-tf does not require to compile the data before the training. Instead, it can directly read text files and preprocess the data when needed by the training. This allows [on-the-fly tokenization](https://opennmt.net/OpenNMT-tf/tokenization.html) and data augmentation by injecting random noise.\n\n### Model fine-tuning\n\nOpenNMT-tf supports model fine-tuning workflows:\n\n* Model weights can be transferred to new word vocabularies, e.g. to inject domain terminology before fine-tuning on in-domain data\n* [Contrastive learning](https://ai.google/research/pubs/pub48253/) to reduce word omission errors\n\n### Source-target alignment\n\nSequence to sequence models can be trained with [guided alignment](https://arxiv.org/abs/1607.01628) and alignment information are returned as part of the translation API.\n\n---\n\nOpenNMT-tf also implements most of the techniques commonly used to train and evaluate sequence models, such as:\n\n* automatic evaluation during the training\n* multiple decoding strategy: greedy search, beam search, random sampling\n* N-best rescoring\n* gradient accumulation\n* scheduled sampling\n* checkpoint averaging\n* ... and more!\n\n*See the [documentation](https://opennmt.net/OpenNMT-tf/) to learn how to use these features.*\n\n## Usage\n\nOpenNMT-tf requires:\n\n* Python 3.7 or above\n* TensorFlow 2.6, 2.7, 2.8, 2.9, 2.10, 2.11, 2.12, or 2.13\n\nWe recommend installing it with `pip`:\n\n```bash\npip install --upgrade pip\npip install OpenNMT-tf\n```\n\n*See the [documentation](https://opennmt.net/OpenNMT-tf/installation.html) for more information.*\n\n### Command line\n\nOpenNMT-tf comes with several command line utilities to prepare data, train, and evaluate models.\n\nFor all tasks involving a model execution, OpenNMT-tf uses a unique entrypoint: `onmt-main`. A typical OpenNMT-tf run consists of 3 elements:\n\n* the **model** type\n* the **parameters** described in a YAML file\n* the **run** type such as `train`, `eval`, `infer`, `export`, `score`, `average_checkpoints`, or `update_vocab`\n\nthat are passed to the main script:\n\n```\nonmt-main --model_type <model> --config <config_file.yml> --auto_config <run_type> <run_options>\n```\n\n*For more information and examples on how to use OpenNMT-tf, please visit [our documentation](https://opennmt.net/OpenNMT-tf).*\n\n### Library\n\nOpenNMT-tf also exposes [well-defined and stable APIs](https://opennmt.net/OpenNMT-tf/package/overview.html), from high-level training utilities to low-level model layers and dataset transformations.\n\nFor example, the `Runner` class can be used to train and evaluate models with few lines of code:\n\n```python\nimport opennmt\n\nconfig = {\n \"model_dir\": \"/data/wmt-ende/checkpoints/\",\n \"data\": {\n \"source_vocabulary\": \"/data/wmt-ende/joint-vocab.txt\",\n \"target_vocabulary\": \"/data/wmt-ende/joint-vocab.txt\",\n \"train_features_file\": \"/data/wmt-ende/train.en\",\n \"train_labels_file\": \"/data/wmt-ende/train.de\",\n \"eval_features_file\": \"/data/wmt-ende/valid.en\",\n \"eval_labels_file\": \"/data/wmt-ende/valid.de\",\n }\n}\n\nmodel = opennmt.models.TransformerBase()\nrunner = opennmt.Runner(model, config, auto_config=True)\nrunner.train(num_devices=2, with_eval=True)\n```\n\nHere is another example using OpenNMT-tf to run efficient beam search with a self-attentional decoder:\n\n```python\ndecoder = opennmt.decoders.SelfAttentionDecoder(num_layers=6, vocab_size=32000)\n\ninitial_state = decoder.initial_state(\n memory=memory, memory_sequence_length=memory_sequence_length\n)\n\nbatch_size = tf.shape(memory)[0]\nstart_ids = tf.fill([batch_size], opennmt.START_OF_SENTENCE_ID)\n\ndecoding_result = decoder.dynamic_decode(\n target_embedding,\n start_ids=start_ids,\n initial_state=initial_state,\n decoding_strategy=opennmt.utils.BeamSearch(4),\n)\n```\n\nMore examples using OpenNMT-tf as a library can be found online:\n\n* The directory [examples/library](https://github.com/OpenNMT/OpenNMT-tf/tree/master/examples/library) contains additional examples that use OpenNMT-tf as a library\n* [nmt-wizard-docker](https://github.com/OpenNMT/nmt-wizard-docker) uses the high-level `opennmt.Runner` API to wrap OpenNMT-tf with a custom interface for training, translating, and serving\n\n*For a complete overview of the APIs, see the [package documentation](https://opennmt.net/OpenNMT-tf/package/overview.html).*\n\n## Additional resources\n\n* [Documentation](https://opennmt.net/OpenNMT-tf)\n* [Forum](https://forum.opennmt.net)\n* [Gitter](https://gitter.im/OpenNMT/OpenNMT-tf)\n\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Neural machine translation and sequence learning using TensorFlow",
"version": "2.32.0",
"project_urls": {
"Documentation": "https://opennmt.net/OpenNMT-tf/",
"Forum": "https://forum.opennmt.net/",
"Gitter": "https://gitter.im/OpenNMT/OpenNMT-tf",
"Homepage": "https://opennmt.net",
"Source": "https://github.com/OpenNMT/OpenNMT-tf/"
},
"split_keywords": [
"tensorflow",
"opennmt",
"nmt",
"neural",
"machine",
"translation"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "ff2db655f288685a9c6b62384d97d2b9132861980e84f8c15af0de48d8a3e5f9",
"md5": "994570a8a71dd5fd184a53fb2bb366d9",
"sha256": "41b4221777fed15a84247123eebba6a1f6becef824a85af33248416a6d331cd5"
},
"downloads": -1,
"filename": "OpenNMT_tf-2.32.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "994570a8a71dd5fd184a53fb2bb366d9",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7",
"size": 162022,
"upload_time": "2023-08-04T08:38:11",
"upload_time_iso_8601": "2023-08-04T08:38:11.131488Z",
"url": "https://files.pythonhosted.org/packages/ff/2d/b655f288685a9c6b62384d97d2b9132861980e84f8c15af0de48d8a3e5f9/OpenNMT_tf-2.32.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "39a7aa003550a746f2c843718b55f21a6697a61b2c9e10587c5ba0fa4b62708b",
"md5": "1c360d336c32891c0cd6e303f5707e64",
"sha256": "dcc7d80046bf6e94f3d7f2e477d1c89c7c57ef6b99f7d9b92235341442833c79"
},
"downloads": -1,
"filename": "OpenNMT-tf-2.32.0.tar.gz",
"has_sig": false,
"md5_digest": "1c360d336c32891c0cd6e303f5707e64",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.7",
"size": 133017,
"upload_time": "2023-08-04T08:38:13",
"upload_time_iso_8601": "2023-08-04T08:38:13.316998Z",
"url": "https://files.pythonhosted.org/packages/39/a7/aa003550a746f2c843718b55f21a6697a61b2c9e10587c5ba0fa4b62708b/OpenNMT-tf-2.32.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-08-04 08:38:13",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "OpenNMT",
"github_project": "OpenNMT-tf",
"travis_ci": false,
"coveralls": true,
"github_actions": true,
"lcname": "opennmt-tf"
}
JSON
