From 16488501bb35f55641b56fca66c7f8f7b34a5af8 Mon Sep 17 00:00:00 2001 From: JS Date: Tue, 12 Mar 2024 15:05:17 +0200 Subject: [PATCH] improve quickstart and tutorial --- _sources/examples/sharing_schemes.md.txt | 7 +- _sources/main.md.txt | 7 +- _sources/prepare_data.md.txt | 80 +++++++---- _sources/quickstart.md.txt | 169 +++++++++++++++------- examples/sharing_schemes.html | 10 +- examples/train_mammoth_101.html | 6 +- index.html | 4 +- main.html | 6 +- modular_model.html | 2 +- prepare_data.html | 78 ++++++---- quickstart.html | 176 +++++++++++++++-------- searchindex.js | 2 +- 12 files changed, 356 insertions(+), 191 deletions(-) diff --git a/_sources/examples/sharing_schemes.md.txt b/_sources/examples/sharing_schemes.md.txt index 9734e79c..1c46373c 100644 --- a/_sources/examples/sharing_schemes.md.txt +++ b/_sources/examples/sharing_schemes.md.txt @@ -117,7 +117,12 @@ This tutorial utilizes SLURM for job scheduling and parallel computing. You can tailor the provided commands for your specific needs, adapting them to alternative job scheduling systems or standalone setups. Ensure that the `config.yaml` file specifies the desired sharing scheme. -#### 3. **Testing Command:** +The training can be run on a single GPU in which case the wrapper wouldn't be necessary. In this case, you can train with the following command. +```bash +python -u $MAMMOTH/train.py -config $CONFIG +``` + +#### 3. **Inference Command:** After training, use the following command to test the model: ```bash diff --git a/_sources/main.md.txt b/_sources/main.md.txt index 702f790c..217bad85 100644 --- a/_sources/main.md.txt +++ b/_sources/main.md.txt @@ -8,10 +8,7 @@ This portal provides a detailed documentation of the **MAMMOTH**: Modular Adapta ## Installation ```bash -git clone https://github.com/Helsinki-NLP/mammoth.git -cd mammoth -pip3 install -e . -pip3 install sentencepiece==0.1.97 sacrebleu==2.3.1 +pip install mammoth-nlp ``` Check out the [installation guide](install) to install in specific clusters. @@ -58,4 +55,4 @@ We published [FoTraNMT](https://github.com/Helsinki-NLP/FoTraNMT) the ancestor o url = "https://aclanthology.org/2023.nodalida-1.24", pages = "238--247" } -``` \ No newline at end of file +``` diff --git a/_sources/prepare_data.md.txt b/_sources/prepare_data.md.txt index 8eaa1dfc..6d467869 100644 --- a/_sources/prepare_data.md.txt +++ b/_sources/prepare_data.md.txt @@ -1,84 +1,108 @@ - # Prepare Data -## UNPC -[UNPC](https://opus.nlpl.eu/UNPC/corpus/version/UNPC) consists of manually translated UN documents from the last 25 years (1990 to 2014) for the six official UN languages, Arabic, Chinese, English, French, Russian, and Spanish. -We preprocess the data. You can download the processed data by: -```bash -wget https://mammoth-share.a3s.fi/unpc.tar -``` -Or you can use the scripts provided by the tarball to process the data yourself. +Before running these scripts, make sure that you have [installed](quickstart) Mamooth, which includes the dependencies required below. -For references, please cite this reference: Ziemski, M., Junczys-Dowmunt, M., and Pouliquen, B., (2016), The United Nations Parallel Corpus, Language Resources and Evaluation (LREC’16), Portorož, Slovenia, May 2016. -## Europarl +## Quickstart: Europarl + +In the [Quickstart tutorial](quickstart), we assume that you will download and preprocess the Europarl data by following the steps below. ### Step 1: Download the data [Europarl parallel corpus](https://www.statmt.org/europarl/) is a multilingual resource extracted from European Parliament proceedings and contains texts in 21 European languages. Download the Release v7 - a further expanded and improved version of the Europarl corpus on 15 May 2012 - from the original website or download the processed data by us: ```bash wget https://mammoth101.a3s.fi/europarl.tar.gz +mkdir europarl_data +tar –xvzf europarl.tar.gz -C europarl_data +``` +Note that the extracted dataset will require around 30GB of memory. Alternatively, you can only download the data for the three example languages (666M). +```bash +wget https://mammoth101.a3s.fi/europarl-3langs.tar.gz +mkdir europarl_data +tar –xvzf europarl-3langs.tar.gz -C europarl_data ``` -We use a SentencePiece model trained on OPUS Tatoeba Challenge data with 64k vocabulary size. Download the SentencePiece model and the vocabulary: +We use a SentencePiece tokenizer trained on OPUS Tatoeba Challenge data with 64k vocabulary size. Download the SentencePiece model and the vocabulary: ```bash # Download the SentencePiece model wget https://mammoth101.a3s.fi/opusTC.mul.64k.spm # Download the vocabulary wget https://mammoth101.a3s.fi/opusTC.mul.vocab.onmt -``` +mkdir vocab +mv opusTC.mul.64k.spm vocab/. +mv opusTC.mul.vocab.onmt vocab/. +``` +If you would like to create and use a custom sentencepiece tokenizer, take a look at the OPUS tutorial below. ### Step 2: Tokenization -Then, read parallel text data, processes it, and generate output files for training and validation sets. +Then, read parallel text data, processes it, and generates output files for training and validation sets. Here's a high-level summary of the main processing steps. For each language in 'langs,' - read parallel data files. - clean the data by removing empty lines. - shuffle the data randomly. - tokenizes the text using SentencePiece and writes the tokenized data to separate output files for training and validation sets. -We use a positional argument 'lang' that can accept one or more values, for specifying the languages (e.g., `bg` and `cs` as used in Europarl) to process. - You're free to skip this step if you directly download the processed data. ```python -import argparse import random +import pathlib import tqdm import sentencepiece as sp -parser = argparse.ArgumentParser() -parser.add_argument('lang', nargs='+') -langs = parser.parse_args().lang +langs = ["bg", "cs"] sp_path = 'vocab/opusTC.mul.64k.spm' spm = sp.SentencePieceProcessor(model_file=sp_path) +input_dir = 'europarl_data/europarl' +output_dir = 'europarl_data/encoded' + for lang in tqdm.tqdm(langs): - en_side_in = f'{lang}-en/europarl-v7.{lang}-en.en' - xx_side_in = f'{lang}-en/europarl-v7.{lang}-en.{lang}' + en_side_in = f'{input_dir}/{lang}-en/europarl-v7.{lang}-en.en' + xx_side_in = f'{input_dir}/{lang}-en/europarl-v7.{lang}-en.{lang}' with open(xx_side_in) as xx_stream, open(en_side_in) as en_stream: data = zip(map(str.strip, xx_stream), map(str.strip, en_stream)) data = [(xx, en) for xx, en in tqdm.tqdm(data, leave=False, desc=f'read {lang}') if xx and en] # drop empty lines random.shuffle(data) - en_side_out = f'{lang}-en/valid.{lang}-en.en.sp' - xx_side_out = f'{lang}-en/valid.{lang}-en.{lang}.sp' + pathlib.Path(output_dir).mkdir(exist_ok=True) + en_side_out = f'{output_dir}/valid.{lang}-en.en.sp' + xx_side_out = f'{output_dir}/valid.{lang}-en.{lang}.sp' with open(xx_side_out, 'w') as xx_stream, open(en_side_out, 'w') as en_stream: for xx, en in tqdm.tqdm(data[:1000], leave=False, desc=f'valid {lang}'): print(*spm.encode(xx, out_type=str), file=xx_stream) print(*spm.encode(en, out_type=str), file=en_stream) - en_side_out = f'{lang}-en/train.{lang}-en.en.sp' - xx_side_out = f'{lang}-en/train.{lang}-en.{lang}.sp' + en_side_out = f'{output_dir}/train.{lang}-en.en.sp' + xx_side_out = f'{output_dir}/train.{lang}-en.{lang}.sp' with open(xx_side_out, 'w') as xx_stream, open(en_side_out, 'w') as en_stream: for xx, en in tqdm.tqdm(data[1000:], leave=False, desc=f'train {lang}'): print(*spm.encode(xx, out_type=str), file=xx_stream) print(*spm.encode(en, out_type=str), file=en_stream) ``` +The script will produce encoded datasets in `europarl_data/encoded` that you can further use for the training. -## OPUS 100 -To get started, download the opus 100 dataset from [OPUS 100](https://opus.nlpl.eu/opus-100.php) +## UNPC +[UNPC](https://opus.nlpl.eu/UNPC/corpus/version/UNPC) consists of manually translated UN documents from the last 25 years (1990 to 2014) for the six official UN languages, Arabic, Chinese, English, French, Russian, and Spanish. +We preprocess the data. You can download the processed data by: +```bash +wget https://mammoth-share.a3s.fi/unpc.tar +``` +Or you can use the scripts provided by the tarball to process the data yourself. + +For references, please cite this reference: Ziemski, M., Junczys-Dowmunt, M., and Pouliquen, B., (2016), The United Nations Parallel Corpus, Language Resources and Evaluation (LREC’16), Portorož, Slovenia, May 2016. + + +## OPUS 100 + +In this guideline, we will also create our custom sentencepiece tokenizer. + +To do that, you will also need to compile a sentencepiece installation in your environment (not just pip install). +Follow the instructions on [sentencepiece github](https://github.com/google/sentencepiece?tab=readme-ov-file#build-and-install-sentencepiece-command-line-tools-from-c-source). + +After that, download the opus 100 dataset from [OPUS 100](https://opus.nlpl.eu/legacy/opus-100.php) ### Step 1: Set relevant paths, variables and download @@ -227,4 +251,4 @@ cd $SP_PATH > $DATA_PATH/zero-shot/$lp/opus.$lp-test.$tl.sp cd $CUR_DIR done -``` \ No newline at end of file +``` diff --git a/_sources/quickstart.md.txt b/_sources/quickstart.md.txt index 8d5d6d85..416c3e82 100644 --- a/_sources/quickstart.md.txt +++ b/_sources/quickstart.md.txt @@ -2,52 +2,82 @@ # Quickstart +MAMMOTH is specifically designed for distributed training of modular systems in multi-GPUs SLURM environments. + +In the example below, we will show you how to configure Mammoth to train a machine translation model with language-specific encoders and decoders. + ### Step 0: Install mammoth ```bash -git clone https://github.com/Helsinki-NLP/mammoth.git -cd mammoth -pip3 install -e . -pip3 install sentencepiece==0.1.97 sacrebleu==2.3.1 +pip install mammoth-nlp ``` Check out the [installation guide](install) to install in specific clusters. ### Step 1: Prepare the data -Prepare the data for training. You can refer to the data preparation [tutorial](prepare_data) for more details. +Before running the training, we will download data for chosen pairs of languages and create a sentencepiece tokenizer for the model. +**Refer to the data preparation [tutorial](prepare_data) for more details.** + +In the following steps, we assume that you already have an encoded dataset containing `*.sp` file for `europarl` dataset, and languages `cs` and `bg`. Thus, your data directory `europarl_data/encoded` should contain 8 files in a format `{train/valid}.{cs/bg}-en.{cs/bg}.sp`. If you use other datasets, please update the paths in the configurations below. ### Step 2: Configurations -You will need to configure your training settings. -Below is a list of configuration examples: + +Mamooth uses configurations to build a new transformer model and configure your training settings, such as which modules are trained with the data from which languages. + +Below are a few examples of training configurations that will work for you out-of-box in a one-node, two-GPU environment.
Task-specific encoders and decoders +In this example, we create a model with encoders and decoders **unshared** for the specified languages. This is defined by `enc_sharing_group` and `enc_sharing_group`. +Note that the configs expect you have access to 2 GPUs. + ```yaml +# TRAINING CONFIG +world_size: 2 +gpu_ranks: [0, 1] + +batch_type: tokens +batch_size: 4096 + +# INPUT/OUTPUT VOCABULARY CONFIG + +src_vocab: + bg: vocab/opusTC.mul.vocab.onmt + cs: vocab/opusTC.mul.vocab.onmt + en: vocab/opusTC.mul.vocab.onmt +tgt_vocab: + cs: vocab/opusTC.mul.vocab.onmt + en: vocab/opusTC.mul.vocab.onmt + +# MODEL CONFIG + +model_dim: 512 + tasks: train_bg-en: src_tgt: bg-en enc_sharing_group: [bg] dec_sharing_group: [en] node_gpu: "0:0" - path_src: /path/to/train.bg-en.bg - path_tgt: /path/to/train.bg-en.en + path_src: europarl_data/encoded/train.bg-en.bg.sp + path_tgt: europarl_data/encoded/train.bg-en.en.sp train_cs-en: src_tgt: cs-en enc_sharing_group: [cs] dec_sharing_group: [en] node_gpu: "0:1" - path_src: /path/to/train.cs-en.cs - path_tgt: /path/to/train.cs-en.en + path_src: europarl_data/encoded/train.cs-en.cs.sp + path_tgt: europarl_data/encoded/train.cs-en.en.sp train_en-cs: src_tgt: en-cs enc_sharing_group: [en] dec_sharing_group: [cs] node_gpu: "0:1" - path_src: /path/to/train.cs-en.en - path_tgt: /path/to/train.cs-en.cs + path_src: europarl_data/encoded/train.cs-en.en.sp + path_tgt: europarl_data/encoded/train.cs-en.cs.sp enc_layers: [6] dec_layers: [6] @@ -58,29 +88,52 @@ dec_layers: [6]
Arbitrarily shared layers in encoders and task-specific decoders +The training and vocab config is the same as in the previous example. + ```yaml +# TRAINING CONFIG +world_size: 2 +gpu_ranks: [0, 1] + +batch_type: tokens +batch_size: 4096 + +# INPUT/OUTPUT VOCABULARY CONFIG + +src_vocab: + bg: vocab/opusTC.mul.vocab.onmt + cs: vocab/opusTC.mul.vocab.onmt + en: vocab/opusTC.mul.vocab.onmt +tgt_vocab: + cs: vocab/opusTC.mul.vocab.onmt + en: vocab/opusTC.mul.vocab.onmt + +# MODEL CONFIG + +model_dim: 512 + tasks: train_bg-en: src_tgt: bg-en enc_sharing_group: [bg, all] dec_sharing_group: [en] node_gpu: "0:0" - path_src: /path/to/train.bg-en.bg - path_tgt: /path/to/train.bg-en.en + path_src: europarl_data/encoded/train.bg-en.bg.sp + path_tgt: europarl_data/encoded/train.bg-en.en.sp train_cs-en: src_tgt: cs-en enc_sharing_group: [cs, all] dec_sharing_group: [en] node_gpu: "0:1" - path_src: /path/to/train.cs-en.cs - path_tgt: /path/to/train.cs-en.en + path_src: europarl_data/encoded/train.cs-en.cs.sp + path_tgt: europarl_data/encoded/train.cs-en.en.sp train_en-cs: src_tgt: en-cs enc_sharing_group: [en, all] dec_sharing_group: [cs] node_gpu: "0:1" - path_src: /path/to/train.cs-en.en - path_tgt: /path/to/train.cs-en.cs + path_src: europarl_data/encoded/train.cs-en.en.sp + path_tgt: europarl_data/encoded/train.cs-en.cs.sp enc_layers: [4, 4] dec_layers: [4] @@ -90,61 +143,71 @@ dec_layers: [4]
Non-modular multilingual system +In this example, we share the input/output vocabulary over all languages. Hence, we define a vocabulary for an `all` language, that we use in the definition of the model. + ```yaml +# TRAINING CONFIG +world_size: 2 +gpu_ranks: [0, 1] + +batch_type: tokens +batch_size: 4096 + +# INPUT/OUTPUT VOCABULARY CONFIG + +src_vocab: + all: vocab/opusTC.mul.vocab.onmt +tgt_vocab: + all: vocab/opusTC.mul.vocab.onmt + +# MODEL CONFIG + +model_dim: 512 + tasks: train_bg-en: src_tgt: all-all - enc_sharing_group: [all] - dec_sharing_group: [all] + enc_sharing_group: [shared_enc] + dec_sharing_group: [shared_dec] node_gpu: "0:0" - path_src: /path/to/train.bg-en.bg - path_tgt: /path/to/train.bg-en.en + path_src: europarl_data/encoded/train.bg-en.bg.sp + path_tgt: europarl_data/encoded/train.bg-en.en.sp train_cs-en: src_tgt: all-all - enc_sharing_group: [all] - dec_sharing_group: [all] + enc_sharing_group: [shared_enc] + dec_sharing_group: [shared_dec] node_gpu: "0:1" - path_src: /path/to/train.cs-en.cs - path_tgt: /path/to/train.cs-en.en + path_src: europarl_data/encoded/train.cs-en.cs.sp + path_tgt: europarl_data/encoded/train.cs-en.en.sp train_en-cs: src_tgt: all-all - enc_sharing_group: [all] - dec_sharing_group: [all] + enc_sharing_group: [shared_enc] + dec_sharing_group: [shared_dec] node_gpu: "0:1" - path_src: /path/to/train.cs-en.en - path_tgt: /path/to/train.cs-en.cs + path_src: europarl_data/encoded/train.cs-en.en.sp + path_tgt: europarl_data/encoded/train.cs-en.cs.sp enc_layers: [6] dec_layers: [6] ```
+**To proceed, copy-paste one of these configurations into a new file named `my_config.yaml`.** -We recommend our [automatic configuration generation tool](config_config) for generating your configurations. +For further information, check out the documentation of all parameters in **[train.py](options/train)**. +For more complex scenarios, we recommend our [automatic configuration generation tool](config_config) for generating your configurations. -### Step 3: Start training +## Step 3: Start training -Now that you've prepared your data and configured the settings, it's time to initiate the training of your multilingual machine translation model using Mammoth. Follow these steps to launch the training script, for example, through the Slurm manager: +You can start your training on a single machine, by simply running a python script `train.py`, possibly with a definition of your desired GPUs. +Note that the example config above assumes two GPUs available on one machine. -```bash -python -u "$@" --node_rank $SLURM_NODEID -u ${PATH_TO_MAMMOTH}/train.py \ - -config ${CONFIG_DIR}/your_config.yml \ - -save_model ${SAVE_DIR}/models/${EXP_ID} \ - -master_port 9974 -master_ip $SLURMD_NODENAME \ - -tensorboard -tensorboard_log_dir ${LOG_DIR}/${EXP_ID} +```shell +CUDA_VISIBLE_DEVICES=0,1 python3 train.py -config my_config.yaml -save_model output_dir -tensorboard -tensorboard_log_dir log_dir ``` -Explanation of Command: - - `python -u "$@"`: Initiates the training script using Python. - - `--node_rank $SLURM_NODEID`: Specifies the node rank using the environment variable provided by Slurm. - - `-u ${PATH_TO_MAMMOTH}/train.py`: Specifies the path to the Mammoth training script. - - `-config ${CONFIG_DIR}/your_config.yml`: Specifies the path to your configuration file. - - `-save_model ${SAVE_DIR}/models/${EXP_ID}`: Defines the directory to save the trained models, incorporating an experiment identifier (`${EXP_ID}`). - - `-master_port 9974 -master_ip $SLURMD_NODENAME`: Sets the master port and IP for communication. - - `-tensorboard -tensorboard_log_dir ${LOG_DIR}/${EXP_ID}`: Enables TensorBoard logging, specifying the directory for TensorBoard logs. - -Your training process has been initiated through the Slurm manager, leveraging the specified configuration settings. Monitor the progress through the provided logging and visualization tools. Adjust parameters as needed for your specific training requirements. You can also run the command on other workstations by modifying the parameters accordingly. +Note that when running `train.py`, you can use all the parameters from [train.py](options/train) as cmd arguments. In the case of duplicate arguments, the cmd parameters override the ones found in your config.yaml. @@ -154,7 +217,7 @@ Now that you have successfully trained your multilingual machine translation mod ```bash python3 -u $MAMMOTH/translate.py \ - --config "${CONFIG_DIR}/your_config.yml" \ + --config "my_config.yml" \ --model "$model_checkpoint" \ --task_id "train_$src_lang-$tgt_lang" \ --src "$path_to_src_language/$lang_pair.$src_lang.sp" \ @@ -166,7 +229,7 @@ python3 -u $MAMMOTH/translate.py \ Follow these configs to translate text with your trained model. - Provide necessary details using the following options: - - Configuration File: `--config "${CONFIG_DIR}/your_config.yml"` + - Configuration File: `--config "my_config.yml"` - Model Checkpoint: `--model "$model_checkpoint"` - Translation Task: `--task_id "train_$src_lang-$tgt_lang"` @@ -182,4 +245,4 @@ Follow these configs to translate text with your trained model. Congratulations! You've successfully translated text using your Mammoth model. Adjust the parameters as needed for your specific translation tasks. ### Further reading -A complete example of training on the Europarl dataset is available at [MAMMOTH101](examples/train_mammoth_101.md), and a complete example for configuring different sharing schemes is available at [MAMMOTH sharing schemes](examples/sharing_schemes.md). \ No newline at end of file +A complete example of training on the Europarl dataset is available at [MAMMOTH101](examples/train_mammoth_101.md), and a complete example for configuring different sharing schemes is available at [MAMMOTH sharing schemes](examples/sharing_schemes.md). diff --git a/examples/sharing_schemes.html b/examples/sharing_schemes.html index c617cc52..d5e764d5 100644 --- a/examples/sharing_schemes.html +++ b/examples/sharing_schemes.html @@ -118,7 +118,7 @@
  • Training Modular Systems
    • -
  • Configure the sharing schemes
    • +
  • MAMMOTH Sharing Schemes

    • Scripts

  • Contributors

    Scripts

    Scripts

    • To proceed, copy-paste one of these configurations into a new file named my_config.yaml.

    +

    For further information, check out the documentation of all parameters in train.py.

    +

    For more complex scenarios, we recommend our automatic configuration generation tool for generating your configurations.

    Step 3: Start training

    -

    Now that you’ve prepared your data and configured the settings, it’s time to initiate the training of your multilingual machine translation model using Mammoth. Follow these steps to launch the training script, for example, through the Slurm manager:

    -
    python -u "$@" --node_rank $SLURM_NODEID -u ${PATH_TO_MAMMOTH}/train.py \
-    -config ${CONFIG_DIR}/your_config.yml \
-    -save_model ${SAVE_DIR}/models/${EXP_ID} \
-    -master_port 9974 -master_ip $SLURMD_NODENAME \
-    -tensorboard -tensorboard_log_dir ${LOG_DIR}/${EXP_ID}
+
    You can start your training on a single machine, by simply running a python script train.py, possibly with a definition of your desired GPUs.
+Note that the example config above assumes two GPUs available on one machine.
    
+
    CUDA_VISIBLE_DEVICES=0,1 python3 train.py -config my_config.yaml -save_model output_dir -tensorboard -tensorboard_log_dir log_dir
 
    
 
    
-
    Explanation of Command:
    
-
      
-
    • python -u "$@": Initiates the training script using Python.
      • 
-
    • --node_rank $SLURM_NODEID: Specifies the node rank using the environment variable provided by Slurm.
      • 
-
    • -u ${PATH_TO_MAMMOTH}/train.py: Specifies the path to the Mammoth training script.
      • 
-
    • -config ${CONFIG_DIR}/your_config.yml: Specifies the path to your configuration file.
      • 
-
    • -save_model ${SAVE_DIR}/models/${EXP_ID}: Defines the directory to save the trained models, incorporating an experiment identifier (${EXP_ID}).
      • 
-
    • -master_port 9974 -master_ip $SLURMD_NODENAME: Sets the master port and IP for communication.
      • 
-
    • -tensorboard -tensorboard_log_dir ${LOG_DIR}/${EXP_ID}: Enables TensorBoard logging, specifying the directory for TensorBoard logs.
      • 
-
    
-
    Your training process has been initiated through the Slurm manager, leveraging the specified configuration settings. Monitor the progress through the provided logging and visualization tools. Adjust parameters as needed for your specific training requirements. You can also run the command on other workstations by modifying the parameters accordingly.
    
-
    +

    Note that when running train.py, you can use all the parameters from train.py as cmd arguments. In the case of duplicate arguments, the cmd parameters override the ones found in your config.yaml.

    -

    Step 4: Translate

    +

    Step 4: Translate

    Now that you have successfully trained your multilingual machine translation model using Mammoth, it’s time to put it to use for translation.

    python3 -u $MAMMOTH/translate.py \
-  --config "${CONFIG_DIR}/your_config.yml" \
+  --config "my_config.yml" \
   --model "$model_checkpoint" \
   --task_id  "train_$src_lang-$tgt_lang" \
   --src "$path_to_src_language/$lang_pair.$src_lang.sp" \
@@ -339,7 +394,7 @@ 
    Step 4: Translate
-
  • Configuration File: --config "${CONFIG_DIR}/your_config.yml"
    • 
+
  • Configuration File: --config "my_config.yml"
    • 
 
  • Model Checkpoint: --model "$model_checkpoint"
    • 
 
  • Translation Task: --task_id "train_$src_lang-$tgt_lang"
    • 
 
@@ -357,9 +412,10 @@ 
    Step 4: Translate
-
    Further reading
    
+
    Further reading
    
 
    A complete example of training on the Europarl dataset is available at MAMMOTH101, and a complete example for configuring different sharing schemes is available at MAMMOTH sharing schemes.
    +
    diff --git a/searchindex.js b/searchindex.js index 4e0febab..46dbfcf7 100644 --- a/searchindex.js +++ b/searchindex.js @@ -1 +1 @@ -Search.setIndex({docnames:["CONTRIBUTING","FAQ","attention_bridges","config_config","examples/sharing_schemes","examples/train_mammoth_101","index","install","main","mammoth","mammoth.inputters","mammoth.modules","mammoth.translate.translation_server","mammoth.translation","modular_model","options/build_vocab","options/server","options/train","options/translate","prepare_data","quickstart","ref"],envversion:{"sphinx.domains.c":1,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":1,"sphinx.domains.index":1,"sphinx.domains.javascript":1,"sphinx.domains.math":2,"sphinx.domains.python":1,"sphinx.domains.rst":1,"sphinx.domains.std":1,"sphinx.ext.viewcode":1,sphinx:56},filenames:["CONTRIBUTING.md","FAQ.md","attention_bridges.md","config_config.md","examples/sharing_schemes.md","examples/train_mammoth_101.md","index.rst","install.md","main.md","mammoth.rst","mammoth.inputters.rst","mammoth.modules.rst","mammoth.translate.translation_server.rst","mammoth.translation.rst","modular_model.md","options/build_vocab.rst","options/server.rst","options/train.rst","options/translate.rst","prepare_data.md","quickstart.md","ref.rst"],objects:{"mammoth.Trainer":{train:[9,1,1,""],validate:[9,1,1,""]},"mammoth.inputters.dataloader":{DynamicDatasetIter:[10,0,1,""],InferenceBatcher:[10,0,1,""],LookAheadBucketing:[10,0,1,""]},"mammoth.inputters.dataloader.DynamicDatasetIter":{from_opts:[10,1,1,""]},"mammoth.inputters.dataloader.LookAheadBucketing":{bucket_is_empty:[10,1,1,""],is_empty:[10,1,1,""],maybe_replenish:[10,1,1,""]},"mammoth.inputters.dataset":{ParallelCorpus:[10,0,1,""]},"mammoth.inputters.vocab":{Vocab:[10,0,1,""]},"mammoth.inputters.vocab.Vocab":{merge:[10,1,1,""]},"mammoth.models":{NMTModel:[9,0,1,""]},"mammoth.models.NMTModel":{count_parameters:[9,1,1,""],forward:[9,1,1,""]},"mammoth.modules":{Embeddings:[11,0,1,""]},"mammoth.modules.Embeddings":{emb_luts:[11,1,1,""],forward:[11,1,1,""],load_pretrained_vectors:[11,1,1,""],word_lut:[11,1,1,""]},"mammoth.modules.attention_bridge":{AttentionBridge:[11,0,1,""]},"mammoth.modules.attention_bridge.AttentionBridge":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.modules.average_attn":{AverageAttention:[11,0,1,""]},"mammoth.modules.average_attn.AverageAttention":{cumulative_average:[11,1,1,""],cumulative_average_mask:[11,1,1,""],forward:[11,1,1,""]},"mammoth.modules.decoder":{DecoderBase:[11,0,1,""]},"mammoth.modules.decoder.DecoderBase":{from_opts:[11,1,1,""]},"mammoth.modules.decoder_ensemble":{EnsembleModel:[11,0,1,""]},"mammoth.modules.encoder":{EncoderBase:[11,0,1,""]},"mammoth.modules.encoder.EncoderBase":{forward:[11,1,1,""]},"mammoth.modules.layer_stack_decoder":{LayerStackDecoder:[11,0,1,""]},"mammoth.modules.layer_stack_decoder.LayerStackDecoder":{add_adapter:[11,1,1,""],forward:[11,1,1,""],from_opts:[11,1,1,""],from_trans_opt:[11,1,1,""],get_submodule:[11,1,1,""],init_state:[11,1,1,""]},"mammoth.modules.layer_stack_encoder":{LayerStackEncoder:[11,0,1,""]},"mammoth.modules.layer_stack_encoder.LayerStackEncoder":{add_adapter:[11,1,1,""],forward:[11,1,1,""],from_opts:[11,1,1,""],from_trans_opt:[11,1,1,""],get_submodule:[11,1,1,""]},"mammoth.modules.mean_encoder":{MeanEncoder:[11,0,1,""]},"mammoth.modules.mean_encoder.MeanEncoder":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.modules.multi_headed_attn":{MultiHeadedAttention:[11,0,1,""]},"mammoth.modules.multi_headed_attn.MultiHeadedAttention":{forward:[11,1,1,""]},"mammoth.modules.position_ffn":{PositionwiseFeedForward:[11,0,1,""]},"mammoth.modules.position_ffn.PositionwiseFeedForward":{forward:[11,1,1,""]},"mammoth.modules.transformer_decoder":{TransformerDecoder:[11,0,1,""]},"mammoth.modules.transformer_decoder.TransformerDecoder":{forward:[11,1,1,""]},"mammoth.modules.transformer_encoder":{TransformerEncoder:[11,0,1,""]},"mammoth.modules.transformer_encoder.TransformerEncoder":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.translate":{BeamSearch:[13,0,1,""],DecodeStrategy:[13,0,1,""],GNMTGlobalScorer:[13,0,1,""],GreedySearch:[13,0,1,""],Translation:[13,0,1,""],TranslationBuilder:[13,0,1,""],Translator:[13,0,1,""]},"mammoth.translate.BeamSearch":{initialize:[13,1,1,""]},"mammoth.translate.DecodeStrategy":{advance:[13,1,1,""],block_ngram_repeats:[13,1,1,""],initialize:[13,1,1,""],maybe_update_forbidden_tokens:[13,1,1,""],maybe_update_target_prefix:[13,1,1,""],target_prefixing:[13,1,1,""],update_finished:[13,1,1,""]},"mammoth.translate.GreedySearch":{advance:[13,1,1,""],initialize:[13,1,1,""],update_finished:[13,1,1,""]},"mammoth.translate.Translation":{log:[13,1,1,""]},"mammoth.translate.Translator":{translate_batch:[13,1,1,""]},"mammoth.translate.greedy_search":{sample_with_temperature:[13,2,1,""]},"mammoth.translate.penalties":{PenaltyBuilder:[13,0,1,""]},"mammoth.translate.penalties.PenaltyBuilder":{coverage_none:[13,1,1,""],coverage_summary:[13,1,1,""],coverage_wu:[13,1,1,""],length_average:[13,1,1,""],length_none:[13,1,1,""],length_wu:[13,1,1,""]},"mammoth.translate.translation_server":{ServerModel:[12,0,1,""],ServerModelError:[12,3,1,""],Timer:[12,0,1,""],TranslationServer:[12,0,1,""]},"mammoth.translate.translation_server.ServerModel":{build_tokenizer:[12,1,1,""],detokenize:[12,1,1,""],do_timeout:[12,1,1,""],maybe_convert_align:[12,1,1,""],maybe_detokenize:[12,1,1,""],maybe_detokenize_with_align:[12,1,1,""],maybe_postprocess:[12,1,1,""],maybe_preprocess:[12,1,1,""],maybe_tokenize:[12,1,1,""],parse_opt:[12,1,1,""],postprocess:[12,1,1,""],preprocess:[12,1,1,""],rebuild_seg_packages:[12,1,1,""],to_gpu:[12,1,1,""],tokenize:[12,1,1,""],tokenizer_marker:[12,1,1,""]},"mammoth.translate.translation_server.TranslationServer":{clone_model:[12,1,1,""],list_models:[12,1,1,""],load_model:[12,1,1,""],preload_model:[12,1,1,""],run:[12,1,1,""],start:[12,1,1,""],unload_model:[12,1,1,""]},"mammoth.utils":{Optimizer:[9,0,1,""],Statistics:[9,0,1,""]},"mammoth.utils.Optimizer":{amp:[9,1,1,""],backward:[9,1,1,""],from_opts:[9,1,1,""],learning_rate:[9,1,1,""],step:[9,1,1,""],training_step:[9,1,1,""],zero_grad:[9,1,1,""]},"mammoth.utils.Statistics":{accuracy:[9,1,1,""],all_gather_stats:[9,1,1,""],all_gather_stats_list:[9,1,1,""],elapsed_time:[9,1,1,""],log_tensorboard:[9,1,1,""],output:[9,1,1,""],ppl:[9,1,1,""],update:[9,1,1,""],xent:[9,1,1,""]},"mammoth.utils.loss":{LossComputeBase:[9,0,1,""]},mammoth:{Trainer:[9,0,1,""]}},objnames:{"0":["py","class","Python class"],"1":["py","method","Python method"],"2":["py","function","Python function"],"3":["py","exception","Python exception"]},objtypes:{"0":"py:class","1":"py:method","2":"py:function","3":"py:exception"},terms:{"1node":5,"24th":8,"25g":7,"4gpu":5,"64k":[5,19],"abstract":11,"boolean":[9,13],"break":[14,19],"byte":17,"case":[4,6],"class":[0,6,9,10,11,12],"default":[12,15,16,17,18,19],"export":[4,5,7],"final":[2,5,11,13],"float":[3,11,13],"function":[0,2,3,4,6,9,11,12,13,17],"import":[0,19],"int":[9,10,11,12,13],"long":0,"new":[0,2],"portoro\u017e":19,"public":7,"return":[0,9,11,12,13],"short":14,"static":[9,17],"t\u00e4ttar":14,"throw":11,"true":[3,5,9,10,11,13,17,18,19],"try":[0,7,10],"while":[3,6,11,14],And:0,Are:6,But:7,EOS:13,For:[0,3,4,5,7,11,13,14,17,19],IDs:13,IFS:19,Its:1,LPs:3,Not:0,One:3,The:[2,4,5,7,9,11,12,13,14,17,19],Then:[0,5,19],There:[2,3],These:[2,3,4,13],Use:[3,17],Used:13,Will:3,__init__:12,_compute_loss:9,a3s:[4,5,19,20],aan:[11,17],aan_useffn:[11,17],ab_fixed_length:17,ab_lay:17,ab_layer_norm:17,abbrevi:0,abil:11,about:[0,6],abov:[0,11,13],abs:[2,17,18,21],acceler:[11,21],accept:[0,3,5,13,19],access:[2,3,7],accord:[3,4],accordingli:20,account:[3,7],accross:9,accum:[9,10],accum_count:[5,9,17],accum_step:[9,17],accumul:[9,17],accuraci:[5,9,13],achiev:[3,11,14],achin:21,acl:[8,21],aclantholog:[8,17],aclweb:17,across:[4,5,14],action:[11,13,17],activ:[1,2,7,11,17],activation_fn:11,activationfunct:11,actual:13,adadelta:17,adafactor:[5,17],adagrad:17,adagrad_accumulator_init:17,adam:17,adam_beta1:17,adam_beta2:17,adamoptim:17,adamw:17,adapt:[1,4,6,8,10,11,14],adapter_group:11,adapter_nam:3,add:[0,10,11],add_adapt:11,add_argu:19,added:3,adding:[0,1],addit:[0,4,11,15,17,18],addition:[4,11],address:[8,13],adjust:[3,4,20],adopt:[1,17],advanc:[13,17],advic:0,aforement:4,after:[0,2,4,13,17,18],afterward:11,again:0,against:11,aidan:21,alessandro:8,alexand:8,algorithm:[6,21],align:[6,9,11,12,13,18],align_debug:18,alignment_head:[11,17],alignment_lay:[11,17],aliv:13,alive_attn:13,alive_seq:13,all:[0,3,4,9,10,11,13,14,15,17,18,20,21],all_gather_stat:9,all_gather_stats_list:9,all_preprocess:12,allennlp:0,alloc:[3,5,14],allow:[0,1,2,3,4,14,17],almost:[13,17],alon:0,alpha:[13,18],alphabet:3,alreadi:[5,15,17,18],also:[0,3,4,5,7,9,11,14,17,20],altern:[4,11],although:11,alwai:[0,3,11],amd:7,among:[4,14],amp:[9,17],anatomi:6,ancestor:8,ancestri:14,ani:[0,3,6,13,15,17,18],anoth:[0,2,9],antholog:17,apex:17,apex_opt_level:17,api:[0,6],api_doc:17,appear:3,append:[7,19],appl:14,appli:[2,3,11,13,15,17,18],applic:18,approach:[6,14],appropri:[4,13],approxim:17,arab:[4,19],arbitrarili:20,architectur:[1,2,14],arg:[0,11,12],argmax:18,argpars:[12,19],argument:[0,5,6,19],argumentpars:19,arn:8,arxiv:[0,2,17,18,21],ashish:21,ask:6,aspect:14,assign:[3,5,18],assum:[11,13],att_typ:2,attend:2,attent:[0,6,9,13,14,18,21],attention_bridg:[9,11],attention_dropout:[11,17],attentionbridg:11,attentionbridgenorm:2,attn:[11,13,18],attn_debug:[13,18],attn_typ:11,attr:12,attribut:[13,14],attributeerror:11,augment:21,author:8,autodoc:0,autogener:17,automat:[1,5,20],avail:[4,9,12,14,17,18,20],available_model:16,averag:[11,17,18,21],average_attn:11,average_decai:[9,17],average_everi:[9,17],average_output:11,averageattent:11,avg:[11,18],avg_raw_prob:18,avoid:[0,3],aws:7,axi:13,back:9,backend:17,backward:9,bahdanau:17,balanc:14,ban_unk_token:[13,18],bank:11,barri:21,bart:[5,15,17,18],base:[0,1,2,3,4,5,7,8,9,10,11,12,13,14,15,17,18,20],baselin:17,basemodel:9,basenam:19,bash:7,batch:[2,5,9,10,11,13,17,18,20],batch_siz:[4,5,10,11,13,17,18,20],batch_size_multipl:[10,17],batch_typ:[5,10,17,18],beam:[6,13],beam_search:13,beam_siz:[13,18],beamsearch:13,beamsearchbas:13,becaus:[3,18],becom:3,been:[13,15,17,18,20],befor:[0,4,12,13,17,18],begin:[9,13],behavior:14,below:[0,4,14,20],ben:3,bengali:3,best:[13,18],beta1:17,beta2:17,beta:[13,18],better:[0,4,7,15,17,18],between:[2,4,6,11,14,15,17,18,21],beyond:9,bia:11,biao:21,bib:0,bibtex:0,bibtext:0,bidir_edg:17,bidirect:17,bin:[7,17],binari:11,bit:18,bitext:14,blank:0,block:[13,18],block_ngram_repeat:[13,18],boggia:8,booktitl:8,bool:[9,10,11,12,13],bos:13,both:[3,4,13,14,17,18],both_embed:[17,18],bound:11,boundari:[15,17,18],bpe:[15,17,18],bptt:[9,17],bridg:[6,21],bridge_extra_nod:17,broad:14,broadcast:14,browser:0,bucket:10,bucket_fn:10,bucket_is_empti:10,buffer:9,build:[0,1,6,9,11,12,13,14,18,19],build_token:12,build_vocab:15,built:[0,9],bulgarian:5,bytetensor:13,cach:11,calcul:[2,9,13],call:[5,11,13],callabl:13,callback:9,can:[2,3,4,5,6,7,9,12,13,14,15,17,18,19,20],cancel:12,candid:[3,15,17,18],cao:21,capabl:[6,14],capit:0,captur:2,care:11,cat:19,catalan:14,categor:13,categori:13,celebr:5,central:14,challeng:[5,14,19],chang:[0,3,9,17],channel:2,charact:0,character:14,character_coverag:19,characterist:14,check:[0,5,8,10,11,20],checklist:0,checkpoint:[4,9,17,20],chen:21,chines:[4,19],chmod:7,choic:[0,10,11,15,17,18],choos:[0,1,4,5,15,17,18],chosen:13,chronopoul:14,citat:[0,6],cite:[0,5,8,19],classmethod:[9,10,11],clean:[5,19],clear:0,clip:5,clone:[8,12,20],clone_model:12,close:0,cls:9,cluster:[3,7,8,14,20],clutter:0,code:[0,3,7],code_dir:7,codebas:[5,7],coder:14,cohes:14,collect:10,column:3,com:[8,20],combin:[14,18],comma:3,command:[6,20],commenc:4,comment:0,common:[0,6,14],commun:[0,1,5,14,20],compar:8,complet:[4,5,13,20],complex:[3,13],compon:[2,3,6,18],composit:17,comput:[3,4,8,9,11,14,17,18],concat:[11,17],conclus:14,condit:[13,17,18],conf:16,confer:8,config:[4,5,6,12,15,16,17,18,20],config_dir:[5,20],config_fil:12,configur:[1,3,4,6,14],congratul:20,connect:2,consid:[1,3,14,19],consider:17,consist:[0,4,19],constant:3,constructor:[0,11],consum:17,contain:[3,4,5,7,11,12,13,14,19],content:[0,18],context:[2,11,17],context_attn:11,continu:0,contribut:[0,2,6],contributor:6,control:[3,9],conv2d:11,conv:11,conveni:[3,4],convent:0,convers:13,convert:[10,12],coordin:14,copi:[0,3,7,11,17,18],copy_attn:[11,13,17],copy_attn_forc:17,copy_attn_typ:17,copy_loss_by_seqlength:17,core:[2,6,9],corpora:[10,14],corpora_info:10,corpu:[3,5,10,14,15,17,18,19],corr:[0,21],correct:[3,4],correspand:12,correspond:[4,10,14,18],could:[13,14],count:[3,9,10,13,15,17,18],count_paramet:9,cov:13,cov_pen:13,coverag:[13,17,18],coverage_attn:17,coverage_non:13,coverage_penalti:[13,18],coverage_summari:13,coverage_wu:13,cpu:[10,12,17,18],crai:7,crayon:17,creat:[3,5,7,9],creation:3,criteria:[14,17],criterion:[5,9],critic:[17,18],cross:[9,11,17],csc:[14,19],csv:3,ct2_model:12,ct2_translate_batch_arg:12,ct2_translator_arg:12,ctrl:0,cuda_visible_devic:5,cumbersom:3,cumul:[11,13,18],cumulative_averag:11,cumulative_average_mask:11,cur_dir:19,cur_len:13,current:[3,9,10,11,13,14,17],curricula:3,curriculum:3,custom:[4,6,12,17],custom_opt:12,customiz:4,cut:[0,19],cutoff:13,d_ff:11,d_model:11,dai:21,data:[1,2,3,4,6,9,13,14,21],data_path:19,data_typ:[9,10,13,17,18],databas:14,dataload:10,datapoint:14,dataset:[6,15,17,18,19,20],dataset_adapt:10,datasetadapt:10,datastructur:12,dblp:0,ddress:21,deal:[3,14],debug:[16,17,18],dec:[3,14],dec_lay:[5,17,20],dec_sharing_group:[4,5,20],decai:17,decay_method:[5,17],decay_step:17,decod:[2,3,6,9,14,20],decode_strategi:13,decoder_ensembl:11,decoder_typ:[5,17],decoderbas:[9,11],decodestrategi:13,deep:14,def:0,defin:[3,5,11,14,15,17,18,20],definit:[11,14],degre:11,delai:3,delet:[15,17,18],delimit:18,delv:4,deng:8,denois:[3,5,6],denoising_object:[15,17,18],denot:[2,4],depend:[0,3,7,9,12],deprec:[17,18],depth:14,desc:19,describ:[2,11,12,17,20],descript:0,design:[4,14],desir:[3,4],detail:[1,4,5,7,8,15,17,18,20],determin:[3,14,18],detoken:12,dev:[7,19],develop:[0,6],devic:[3,4,10,11,13,14,18],device_context:9,deyi:21,diagon:3,diagram:11,dict:[3,9,10,12,13,15,17,18],dict_kei:17,dictionari:[9,11,13,17],differ:[0,1,2,3,4,11,12,14,18,20],dim:11,dimens:[5,11,13,17],dimension:2,dir:19,direct:[0,3,8,13,14],directli:[0,5,18,19],directori:[3,4,7,12,17,20],disabl:17,discard:[15,17,18],discourag:17,disk:[17,18],displai:9,dist:9,distanc:[11,17],distinct:[4,14],distribut:[3,5,9,11,13,14,15,17,18],dive:4,divers:[15,17,18],divid:[2,3,17,18],divis:11,dmon:5,do_timeout:12,doc:0,document:[0,4,6,8,19],doe:[3,6,18],doesn:19,doi:8,doing:[3,18],domain:6,don:0,done:[5,13,14,19],dot:[2,11,17],dotprod:17,dowmunt:19,down:[13,14,15],download:[4,6,7],dozen:8,drop:19,dropout:[5,9,11,15,17,18],dropout_step:[9,17],due:17,dummi:[5,11],dump:[15,17,18],dump_beam:[13,18],dump_sampl:15,dump_transform:[17,18],dure:[11,12,14,17,18],dynam:[6,10,11,18],dynamicdatasetit:10,each:[3,4,5,11,13,14,15,17,18,19],earli:[5,17],earlier:[2,15,17,18],early_stop:[5,17],early_stopping_criteria:[5,17],earlystopp:9,eas:[3,14],easi:0,easili:3,echo:[5,19],ecosystem:1,edg:17,effect:[2,3,12,15],effici:[6,9,14,21],eight:[],either:[13,17],elaps:9,elapsed_tim:9,element:[2,3],els:19,emb_fil:11,emb_lut:11,embed:[2,6,14,15],embedding_s:11,embeddingless:17,embeddings_typ:[17,18],emerg:2,emploi:[2,4,9],empti:[5,10,11,13,15,17,18,19],en_side_in:19,en_side_out:19,en_stream:19,enabl:[1,5,11,14,17,18,20],enable_embeddingless:[11,17],enc:[3,14],enc_hidden:11,enc_lay:[5,17,20],enc_output:11,enc_sharing_group:[4,5,20],encod:[2,3,6,9,13,14,19,20],encoder_typ:[5,17],encoderbas:[9,11],encordec:[15,17,18],encount:[15,17,18],encout:[10,15,17,18],end:13,eng:3,english:[3,4,5,19],enhanc:[2,6,14],ensembl:18,ensemblemodel:11,ensur:[2,4,14],entir:[4,19],entri:0,entropi:9,env_dir:7,environ:[4,7,20],eos:13,epoch:17,epsilon:17,equal:[13,17],equat:11,equival:17,error:[0,11,15,17,18],especi:3,essenti:13,establish:2,etal:8,etc:5,eural:21,europarl:[5,6,20],european:[5,19],evalu:[9,19],even:3,event:13,everi:[9,11,17,18],everyth:14,exactli:0,exampl:[0,3,4,5,7,10,11,14,15,17,18,20],exce:17,except:[0,12,15,17,18],exclusion_token:13,execut:[4,14,15,17,18],exist:[5,11,15,17,18,19],exp:17,exp_host:17,exp_id:[5,20],expand:[5,19],expans:14,expect:[3,13],experi:[4,5,15,17,18,20],experiment:17,expert:1,explan:20,explicitli:14,explor:4,exponenti:17,extend:0,extern:0,extra:[7,11,17],extract:[5,19],facilit:[2,14],fail:13,fairseq:0,fals:[5,9,10,11,12,13,15,16,17,18,19],familiar:8,faro:8,faster:17,favorit:5,feat_0:18,feat_1:18,feat_dim_expon:11,feat_merg:[11,17],feat_merge_s:17,feat_padding_idx:11,feat_vec_expon:[11,17],feat_vec_s:[11,17],feat_vocab_s:11,feats0:18,feats1:18,featur:[2,6,9,11,14,15,18,21],fed:2,feed:[3,11,14,17],feedforward:[2,17],feedforwardattentionbridgelay:6,feel:[0,4],few:0,ffn:[11,17],field:10,fifth:5,figur:[11,14],file:[0,3,4,5,10,12,15,17,18,19,20],filenam:17,filter:[5,6,19],filterfeat:[15,17,18],filternonzeronumer:[15,17,18],filterrepetit:[15,17,18],filterterminalpunct:[15,17,18],filtertoolong:[3,5,15,17,18],filterwordratio:[15,17,18],find:[0,5],finer:14,firefox:0,first:[0,3,5,7,11,13,17],five:2,fix:[0,13,17],flag:[3,4,9],flake8:0,flexibl:[1,4],floattensor:[9,11,13],flow:[2,14],fnn:11,focu:[0,2],focus:14,folder:0,follow:[0,2,3,4,18,20],foo:0,forbidden:13,forbidden_token:13,forc:[13,18],format:[0,12,15,17,18,19],former:11,forward:[3,9,11,14,17],fotranmt:8,found:19,foundat:[2,4],four:[],fp16:[17,18],fp32:[9,17,18],frac:2,fraction:[15,17,18],framework:[4,6,17],free:[0,4,5,12,19],freez:[11,17],freeze_word_vec:11,freeze_word_vecs_dec:17,freeze_word_vecs_enc:17,french:[4,19],frequenc:[15,17,18],frequent:6,friendli:1,from:[1,2,3,5,7,9,10,11,13,14,17,18,19],from_opt:[9,10,11],from_trans_opt:11,from_xxx:3,frozenset:13,full:[0,3,11,12,14,15,17,18,19],full_context_align:[11,17],fulli:[1,3,11,14],further:[5,6,15,17,18,19],fusedadam:17,futur:[6,7],gao:21,gap:21,garg:17,gather:9,gating_output:11,gelu:17,gener:[0,1,2,3,4,5,6,9,10,13,18,19,20],generator_funct:17,ger:14,get:[6,7],get_submodul:11,git:[8,20],github:[8,17,20],give:[3,17,18],given:[2,3,11,12],global_attent:17,global_attention_funct:17,global_scor:13,glove:[17,18],gnmt:13,gnmtglobalscor:13,going:13,gold:13,gold_scor:13,gold_sent:13,gomez:21,gone:17,good:[0,17],googl:[0,13,18,21],gpu:[3,4,5,7,12,13,14,17,18,20],gpu_backend:17,gpu_load:5,gpu_rank:[5,17],gpu_verbose_level:[9,17],gpuid:17,grad:9,gradient:[5,9,14,17],graham:21,grain:14,gram:13,graph:17,gre:7,greater:13,greedy_search:13,greedysearch:13,group:[4,5,11,14,17,18],groupwis:[3,14],grow:13,gtx1080:18,guid:[4,5,8,11,20],guidelin:6,guillaum:8,hack:14,had:18,haddow:21,hand:[3,14],handl:[0,9,14],happen:13,hardwar:14,has:[2,3,4,6,11,13,14,15,17,18,20],has_cov_pen:13,has_len_pen:13,has_tgt:13,have:[0,3,11,13,14,17,18,20],head:[2,5,11,17],head_count:11,help:[0,2,18],helsinki:[8,20],here:[4,5,7,13,19],hidden:[9,11,17],hidden_ab_s:17,hidden_dim:2,hieu:21,high:[3,5,19],higher:[13,17,18],highest:18,hold:13,hook:11,hoorai:5,host:7,hour:5,how:[0,6,11],howev:[0,9],html:[0,17],http:[2,4,5,7,8,17,18,19,20,21],huge:17,human:[3,21],hyp:[4,20],hyper:1,hyphen:3,identifi:[5,18,20],idl:3,ids:3,ignor:[5,11,15,17,18],ignore_when_block:[13,18],iii:14,illia:21,illustr:14,ilya:21,imag:9,impact:17,implement:[2,9,14,17],improv:[1,5,11,13,17,19,21],in_config:3,in_featur:11,includ:[0,1,3,4,5,11,14,15,17,18],incompat:[15,17,18],incorpor:[17,20],increas:[3,14],independ:14,index:[7,11,17],indic:[4,5,9,11,13,15,17,18],individu:[3,11],inf:13,infer:[10,13,14],inferencebatch:10,inferfeat:6,info:[5,10,17,18],inform:[2,3,5,14,17,18],infrastructur:14,ingredi:13,init:17,init_st:[9,11],initi:[4,5,6,9,11,12,13,20],initial_accumulator_valu:17,initil:10,inner:11,innov:6,inp:13,inp_seq_len:13,inproceed:8,input:[2,6,9,10,11,12,13,15,17,18,19,21],input_len:11,input_sentence_s:19,inputs_len:11,inputt:[10,13],insert:[15,17,18],insert_ratio:[15,17,18],insid:5,instal:[0,5,6],instanc:[9,11,13],instanti:9,instead:[0,3,7,11,15,17,18],instruct:[7,17],int8:18,integ:13,integr:0,interact:7,interfac:[9,11],intermedi:2,intermediate_output:2,intern:12,interv:17,introduc:[2,3,6],introduct:3,invalid:[11,15,17,18],involv:[2,14],is_empti:10,is_finish:13,is_normform:11,is_on_top:11,is_train:10,island:8,isn:13,issu:6,item:[10,11],iter:[9,10],iterabledataset:10,its:[0,3,6,14],itself:[3,11],jakob:21,jean:8,jinsong:21,job:[4,7],johnson:14,joiner:[15,17,18],jone:21,journal:0,json:16,junczi:19,kaiser:21,keep:[12,13,17],keep_checkpoint:[5,17],keep_stat:17,keep_topk:13,keep_topp:13,kei:[6,11,14],kera:17,kernel_s:11,key_len:11,kim:8,klau:21,klein:8,koehn:5,krikun:21,kwarg:11,label:[5,17],label_smooth:[5,17],lakew:14,lambda:[15,17,18],lambda_align:17,lambda_coverag:17,lang:[3,5,19],lang_a:3,lang_b:3,lang_pair:[3,20],languag:[2,4,5,14,15,17,18,19,20],language_pair:19,larg:14,last:[3,4,17,18,19],latter:11,launch:20,layer:[2,5,11,14,17,18,20],layer_cach:11,layer_norm_modul:11,layer_stack_decod:11,layer_stack_encod:11,layer_stack_index:11,layer_type_to_cl:2,layernorm:17,layerstack:3,layerstackdecod:11,layerstackencod:11,layerwis:14,lead:13,learn:[2,5,9,14,17],learning_r:[5,9,17],learning_rate_decai:17,learning_rate_decay_fn:9,least:0,leav:[3,17,19],len:[9,11,13],length:[3,9,11,13,15,17,18,19],length_averag:13,length_non:13,length_pen:13,length_penalti:[13,18],length_wu:13,less:3,let:[3,4,11,14],level:[5,6,10,15,17,18,19],leverag:[6,14,20],lib:7,librari:[8,17],lightweight:14,like:[0,4,11,13,14,18],limit:18,lin:[2,14,17],linattentionbridgelay:6,line:[0,5,10,15,17,18,19],linear:[2,11],linear_warmup:17,linguist:[8,11,14,21],link:[0,2,7],list:[0,3,9,11,12,13,15,17,18,20],list_model:12,literatur:17,llion:21,load:[6,7,9,11,12,14,17,18],load_model:12,load_pretrained_vector:11,loader:6,local:[0,3],localhost:17,locat:5,log:[5,6,9,13,20],log_dir:[5,20],log_fil:[17,18],log_file_level:[17,18],log_prob:13,log_tensorboard:9,logger:13,login:7,logit:[13,18],logsumexp:13,longer:18,longest:13,longitudin:14,longtensor:[9,11,13],look:[0,8,10,11,18],look_ahead_s:10,lookaheadbucket:10,loop:9,loppi:8,loss:[6,17],loss_scal:17,losscomputebas:9,love:[0,1],low:14,lower:[3,17,18],lrec:19,lsl:[13,21],lstm:17,lua:12,lukasz:21,lumi:6,luong:[17,21],lust:7,lustrep1:7,lustrep2:7,macherei:21,machin:[1,5,8,11,13,14,17,18,20,21],made:[3,5,6],magic:13,mahti:6,mai:[3,5,8,9,10,12,13,15,17,19],main:[0,5,8,9,15,17,18,19],maintain:[4,13],make:[0,4,7,9,10,14,15,17,18],make_shard_st:9,mammoth101:[5,19,20],mammoth:[0,3,6,7,8,9,10,11,12,13,14,17,19],mammoth_transform:[15,17,18],manag:[4,5,9,14,20],mani:[9,13,17],manipul:9,manual:[4,5,12,13,19],map:[3,9,19],marian:17,mark:17,marker:12,mask:[11,15,17,18],mask_length:[15,17,18],mask_or_step:11,mask_ratio:[5,15,17,18],mass:[15,17,18],massiv:[1,3,8],master:[5,17,20],master_ip:[4,17,20],master_port:[4,5,17,20],match:12,mathbb:2,mathbf:2,matric:2,matrix:[2,11,17],max:[9,11,13,19],max_generator_batch:[5,17],max_grad_norm:[5,9,17],max_length:[13,18],max_relative_posit:[11,17],max_sent_length:18,max_sentence_length:19,max_siz:9,maxim:21,maximum:[15,17,18],maybe_convert_align:12,maybe_detoken:12,maybe_detokenize_with_align:12,maybe_postprocess:12,maybe_preprocess:12,maybe_replenish:10,maybe_token:12,maybe_update_forbidden_token:13,maybe_update_target_prefix:13,mean:[3,4,11,12,14,17,18],mean_encod:11,meanencod:11,mechan:[1,2,3,14],mem:7,memori:[11,12,17],memory_bank:[11,13],memory_length:11,merg:[10,11,17],meta:3,metadata:9,method:[9,11,17],metric:18,mi250:7,michel:8,micku:8,mike:21,million:8,min_length:[13,18],minh:21,minim:14,minimum:[15,17,18],mirror:17,mix:9,mixer:10,mixingstrategi:10,mixtur:1,mkdir:[5,7,19],mlp:[11,17],mnmt:14,mode:[3,15,17,18],model:[1,2,3,4,6,11,13,15,20],model_checkpoint:20,model_dim:[5,11,17],model_dtyp:[9,17],model_fil:19,model_id:12,model_kwarg:12,model_prefix:19,model_root:12,model_sav:9,model_task:17,model_typ:[5,17],modelsaverbas:9,modif:[6,9],modifi:[0,4,13,20],modul:[0,2,6,7,9,14,17],modular:[1,6,8,20],module_id:11,mohammad:21,moment:5,monitor:[5,20],monolingu:3,month:8,more:[0,1,3,5,7,10,13,14,15,17,18,19,20],most:[13,14,18],mostli:9,move:[12,17],moving_averag:[9,17],much:17,mul:[5,19],multi:[0,2,5,11],multi_headed_attn:11,multiheadedattent:[2,11],multilingu:[1,3,5,8,19,20],multinod:5,multipl:[0,2,3,9,10,11,14,17,18],multipli:[2,10],multplic:0,must:[3,11,12,14,17],mymodul:7,n_batch:9,n_best:[12,13,18],n_bucket:[10,17],n_correct:9,n_edge_typ:17,n_node:[5,17],n_sampl:[15,17,18],n_seg:12,n_src_word:9,n_step:17,n_word:9,naacl:17,name:[0,3,5,6,11,13,15,17,18,19],named_modul:11,namespac:12,napoleon:0,narg:19,nation:19,nccl:17,necessari:[0,4,5,7,9,13,17,18,20],necessit:3,need:[0,3,4,5,9,11,14,17,20,21],neg:[12,17],nest:11,net_b:11,net_c:11,network:[1,11,21],neubig:21,neural:[1,8,11,13,14,21],never:13,next:[3,9,13,18],nfeat:11,ngram:[13,18],nightmar:3,niki:[8,21],nlp:[8,20],nmt:[9,13,17,18],nmtmodel:[9,11],noam:[17,21],noamwd:17,nodalida:8,node:[3,5,7,9,14,17,20],node_gpu:[5,20],node_rank:[4,5,17,20],nois:3,non:[11,13,17,20],none:[5,9,10,11,12,13,15,17,18],nonetyp:13,nonzero_threshold:[15,17,18],nordic:8,norm:[11,17],norm_method:9,normal:[2,3,5,9,11,17],normalz:9,normform:[11,17],norouzi:21,note:[0,3,7,13],noth:[0,9],notset:[17,18],now:[1,4,5,20],nroo:8,ntask:7,nucleu:18,num_lay:11,num_step:9,num_thread:15,number:[3,9,10,11,13,14,15,17,18],numel_fn:10,numer:[15,17,18],nvidia:[5,7,17],obj:[0,9],object:[0,9,10,12,13,15,17,18,19],occur:4,oder:3,off:17,offer:[4,14],offici:[4,19],offset:10,ofi:7,often:[15,17,18],on_timemout:12,on_timeout:12,onc:[13,17],one:[0,2,3,4,5,9,10,11,15,17,18,19],onli:[3,4,9,10,13,14,15,17,18],onmt:[5,19],onmt_token:[15,17,18],onmttok:6,open:[1,8,19],opennmt:[0,3,6,8,9,16],oper:14,operatornam:2,opt:[5,9,10,11,12,17,18],opt_level:17,optim:[1,5,6,14],option:[0,3,4,5,7,9,10,11,12,13,15,17,18,19,20],opu:[5,6],opus100:19,opustc:[5,19],ord:21,order:[3,17],org:[2,7,8,17,18,21],organ:[10,14],origin:[5,6,17,19],oriol:21,other:[2,4,5,7,9,13,15,17,18,19,20,21],other_lang:19,otherwis:[3,11,17,18],our:[5,7,13,20],our_stat:9,out:[3,8,9,20],out_config:3,out_featur:11,out_fil:13,out_path:[4,20],out_typ:19,output:[2,3,4,5,9,11,12,13,15,17,18,19,20],output_model:18,over:[0,3,9,13,17,18,19],overal:2,overcom:14,overhead:14,overrid:[11,13,15,17,18],overridden:11,overse:14,overview:6,overwrit:[5,7,15,17,18],own:[6,9,14],ownership:9,p17:8,p18:17,packag:[5,7,12],pad:[9,11,13],page:[5,8],pair:[3,4,9,12,14,17,19],paper:[0,2,5,17],parallel:[4,5,6,11,13,15,17,19],parallel_path:13,parallelcorpu:[10,13],param:9,param_init:[5,17],param_init_glorot:[5,17],paramet:[1,4,5,6,8,9,10,11,12,13,15,17,18,20],parenthes:0,parliament:[5,19],parmar:21,pars:[10,12],parse_arg:19,parse_opt:12,parser:19,part:[2,13],partial:14,particular:[0,3],particularli:6,partit:7,pass:[2,3,9,11,12,17],past:[0,17],path:[3,4,5,7,10,11,12,13,15,17,18,20],path_src:[5,20],path_tgt:[5,20],path_to_codebas:5,path_to_europarl:5,path_to_mammoth:20,path_to_src_languag:20,path_to_vocab:5,path_to_your_env:5,path_valid_src:5,path_valid_tgt:5,patienc:[5,9],pattern:[3,15,17,18],pdf:17,pen:13,penalti:[6,13,15,17],penaltybuild:13,peopl:7,per:[0,3,15,17,18],perceiv:[2,17],perceiverattentionbridgelay:6,percentag:[15,17,18],perfom:17,perform:[2,11,17],permut:[15,17,18],permute_sent_ratio:[15,17,18],perplex:9,pertain:14,pfs:7,pham:21,philipp:5,phrase_t:[13,18],phylogenet:14,pip3:[7,8,20],pip:[0,7],pipelin:[15,17,18],plain:10,plan:6,platform:7,pleas:[0,5,8,19],plu:17,point:[5,20],pointer:5,poisson:[15,17,18],poisson_lambda:[15,17,18],polosukhin:21,pool:[11,17],pool_siz:[10,17],port:[5,16,17,20],portal:8,pos_ffn_activation_fn:[11,17],posit:[5,11,17,19],position_encod:[5,11,17],position_ffn:11,positionalencod:11,positionwisefeedforward:[11,17],possibl:[3,9,11,12,13,15,17,18],postprocess:12,postprocess_opt:12,potenti:13,pouliquen:19,pouta:19,ppid:5,ppl:9,pre:[9,12,13],pre_word_vecs_dec:17,pre_word_vecs_enc:17,preced:3,precis:9,pred:18,pred_scor:13,pred_sent:13,predict:[9,13,18],prefer:0,prefix:[3,9,15,17,18],prefix_seq_len:13,preload:12,preload_model:12,prepar:[6,13],preprint:21,preprocess:[4,12,14,19],preprocess_opt:12,presenc:3,present:5,presum:13,pretrain:[11,17],prevent:[13,18],previou:[3,11,13],previous:2,primari:3,prime:2,print:[9,17,18,19],prior_token:[15,17,18],priori:14,prob:13,proba:18,probabl:[11,13,15,17,18],problem:[13,14],proc:[8,21],procedur:3,proceed:[5,8,19],process:[2,4,5,9,12,14,15,17,19,20],processed_data:4,processu:12,produc:[2,13,15,17,18],product:2,progress:[5,20],projappl:7,project:[0,2,5,6,7,8],project_462000125:7,propag:9,proper:12,properli:7,properti:[9,11],proport:[3,15,17,18],provid:[4,5,7,8,18,19,20],prune:6,pty:7,publish:8,puhti:6,pull_request_chk:0,punct_threshold:[15,17,18],punctuat:[0,15,17,18],purason:14,put:[13,20],pwd:19,pyonmttok:[15,17,18],python3:[3,4,7,20],python:[0,3,4,5,7,17,20],pythonpath:7,pythonuserbas:[5,7],pytorch:[0,1,7],qin:21,qualifi:11,quantiz:18,queri:11,query_len:11,question:6,queue:[15,17],queue_siz:17,quickstart:[6,8],quoc:21,quot:0,raganato:8,rais:[11,15,17,18],random:[6,15,17,19],random_ratio:[15,17,18],random_sampling_temp:[13,18],random_sampling_topk:[13,18],random_sampling_topp:[13,18],randomli:[5,13,19],rang:18,rank:[5,13,14,17,20],ranslat:21,rare:13,rate:[5,6,9],rather:0,ratio:[13,15,17,18],raw:[10,13,18],raw_prob:11,rccl:7,reach:13,read:[0,3,5,6,12,19],readabl:[0,3,17,18],readm:17,real:11,rebuild:12,rebuild_seg_packag:12,receiv:3,recent:17,recip:11,recommend:[4,7,17,20],recommonmark:0,rectifi:2,recurr:11,reduc:14,redund:3,ref:0,refer:[0,1,2,4,6,7,11,19,20],referenc:11,regardless:[3,4],regist:11,regular:[15,17,18],rel:[11,14,17],relat:[5,14,15,17,18],relationship:[2,6],releas:[5,7,19],relev:[11,13,14],relu:[2,11,17],rememb:[0,4],remov:[3,5,15,17,18,19],renorm:17,reorder:13,rep_max_len:[15,17,18],rep_min_len:[15,17,18],rep_threshold:[15,17,18],repeat:[13,15,17,18],repetit:18,replac:[4,13,15,17,18],replace_length:[5,15,17,18],replace_unk:[13,18],replic:11,report:[6,8,9,17,18],report_align:[13,18],report_everi:[5,17],report_manag:9,report_scor:13,report_stats_from_paramet:[9,17],report_tim:[13,18],reportmgrbas:9,repositori:6,repres:[5,9],represent:[2,11,14,17],reproduc:6,requir:[0,4,9,11,14,17,20],research:[5,8],reservoir:10,reset:9,reset_optim:17,resett:17,residu:[11,14],resolv:11,resourc:[3,5,14,19],respect:[2,3],respons:9,rest:16,restrict:[15,17,18],result:[11,12,17],return_attent:13,reus:17,reuse_copy_attn:17,revers:[15,17,18],reversible_token:[15,17,18],rico:21,right:0,rmsnorm:17,rnn:[9,17],roblem:21,rocm5:7,rocm:7,root:[2,3],rotat:[15,17,18],rotate_ratio:[15,17,18],roundrobin:17,row:3,rshavn:8,rsqrt:17,rst:0,run:[0,3,5,9,11,12,17,18,20],runtim:11,rush:8,russian:[4,19],s_idx:10,sacrebleu:[7,8,20],sai:[3,11,14],samantao:7,same:[0,3,11,12,14,17],sampl:[6,13,15,17,19],sample_with_temperatur:13,sampling_temp:13,save:[5,9,14,15,17,18,19,20],save_all_gpu:[5,17],save_checkpoint_step:[5,9,17],save_config:[15,17,18],save_data:[15,17,18],save_dir:[5,20],save_model:[5,17,20],saver:9,scalabl:[1,14],scale:[11,13,14,17],scenario:1,schedul:[4,9,17],scheme:[5,6,14,20],schuster:21,score:[6,12,15,17,18],scorer:13,scratch:[1,7],script:[0,4,5,6,7,19,20],script_dir:5,seamless:4,search:[0,3,6,13],second:[2,3,11,12],section:[5,7],secur:[10,15,17,18],see:[3,5,11,12,13,15,17,18],seed:[5,13,15,17,18],seemingli:17,seen:2,segment:[3,12,15,17,18],select:[11,13],select_index:13,selector:3,self:[2,11,12,13,17],self_attn_typ:[11,17],send:[0,17],senellart:8,sennrich:21,sensibl:0,sent:[9,10,17,18],sent_numb:13,sentenc:[13,15,17,18,19],sentencepiec:[3,6,7,8,15,17,18,20],sentencepieceprocessor:19,separ:[3,5,11,14,19],seper:12,seq2seq:[13,17],seq:13,seq_len:[2,13],sequenc:[1,2,3,9,11,12,13,14,15,17,18],seri:14,serial:11,serv:[2,4],server:[6,17,19],servermodel:12,servermodelerror:12,session:7,set:[1,2,3,4,5,7,9,11,12,13,14,15,17,18,20],sever:[3,5,11,13],sgd:17,sh16:[11,21],shaham:17,shallow:14,shape:[0,11,13],shard:[9,17,18],shard_siz:[4,9,18,20],share:[1,5,6,7,8,11,15,17,18,19,20],share_decoder_embed:17,share_embed:17,share_vocab:[15,17,18],shazeer:21,shortest:13,shot:3,should:[3,5,11,13,17,18],show:11,shuf:19,shuffl:[5,19],shuffle_input_sent:19,side:[3,9,12,15,17,18],side_a:3,side_b:3,sign:[15,17,18],signifi:4,silent:[5,11,15,17,18],similar:[2,3,11,17],simpl:[2,9,11,17],simpleattentionbridgelay:6,simpli:11,simulatan:11,sin:17,sinc:11,singl:[0,5,12,14,17],single_pass:17,singular:7,site:7,six:[4,19],size:[3,5,9,10,11,13,15,17,18,19,20],skip:[3,5,15,17,18,19],skip_embed:11,skip_empty_level:[5,10,15,17,18],slovenia:19,slow:[15,18],slurm:[3,4,5,7,20],slurm_nodeid:[4,5,20],slurmd_nodenam:[4,20],smaller:[14,15,17,18],smi:5,smooth:[5,15,17,18],softmax:[2,17,18],solid:4,some:[0,3,9,11,18],someth:[0,11],sometim:0,soon:5,sort:[12,19],sorted_pair:3,sourc:[0,1,3,4,5,7,8,9,10,11,12,13,14,15,17,18,20],sp_path:19,space:[0,2,14,17],spacer:[15,17,18],span:[4,15,17,18],spanish:[4,19],spars:[1,11],special:10,specif:[2,3,4,5,6,8,13,14,15,17,18,20],specifi:[4,5,11,15,17,18,19,20],sphinx:0,sphinx_rtd_them:0,sphinxcontrib:0,spill:0,spm:[5,19],spm_encod:19,spm_train:19,sqrt:2,squar:[2,3],src:[3,4,9,11,12,13,15,17,18,19,20],src_embed:[17,18],src_feat:18,src_feats_vocab:[15,17,18],src_file:10,src_file_path:13,src_ggnn_size:17,src_group:3,src_lang:[3,20],src_languag:3,src_len:[9,11],src_length:13,src_map:13,src_onmttok_kwarg:[15,17,18],src_raw:13,src_seq_length:[5,15,17,18],src_seq_length_trunc:[17,18],src_subword_alpha:[15,17,18],src_subword_model:[15,17,18],src_subword_nbest:[15,17,18],src_subword_typ:[5,15,17,18],src_subword_vocab:[15,17,18],src_tgt:[5,20],src_vocab:[5,10,13,15,17,18],src_vocab_s:[5,17,18],src_vocab_threshold:[15,17,18],src_words_min_frequ:[17,18],srun:[4,5,7],stabl:2,stack:[11,14,17],stage:2,stand:0,standalon:4,standard:[11,17,18],start:[3,4,5,6,7,9,12,17,19],start_decay_step:17,stat:[9,17],stat_list:9,state:[9,11,13,14,17],state_dict:17,state_dim:17,statist:[5,9,17],statu:6,stdout:9,step:[2,3,6,9,11,13,17,18],stepwis:11,stepwise_penalti:[13,18],stig:8,still:0,stop:[5,15,17,18],store:17,str:[0,9,10,11,12,13,19],strategi:[1,6,9,10,14,17],streamlin:14,stride:[10,11],string:[9,11,15,17,18],strip:19,structur:[2,4,6,17,18],structured_log_fil:[17,18],style:[0,10,11,14,15,17,18],styleguid:0,sub_id:11,subclass:[9,11,13],subcompon:3,subdirectori:7,sublay:6,submodul:11,subsequ:2,subset:19,substitut:3,substr:[15,17,18],subword:[3,6,14],successfulli:20,suggest:17,suit:[4,6],sum:[9,11,13,17],sume:9,summari:[0,5,13,18,19],summit:5,superclass:0,supercomput:7,supervis:[3,11,17],support:[0,1,3,17],suppos:19,sure:[4,7,13],sutskev:21,swahili:14,switchout:[6,21],switchout_temperatur:[15,17,18],symmetr:3,sync:14,synchron:14,system:[6,13,14,17,20,21],t_idx:10,tab:[15,17,18],tabl:[11,18],tag:10,tailor:4,take:[2,3,5,8,11,15,17,18],tar:[4,5,19,20],tarbal:19,target:[3,4,5,9,11,12,13,14,15,17,18],target_prefix:13,tartu:8,task2gpu:14,task:[3,6,9,10,11,13,14,20],task_distribution_strategi:17,task_id:[4,18,20],task_queue_manag:[9,10,11],tatoeba:[3,5,19],tau:[15,17,18],team:7,technic:8,temperatur:[3,13,15,17,18],templat:3,tensor:[0,9,10,11,13],tensorboard:[5,9,17,20],tensorboard_log_dir:[5,17,20],tensorflow:17,termin:[15,17,18],test:[0,1,7],text:[5,9,10,13,17,18,19,20],tgt:[3,9,11,12,15,17,18],tgt_embed:[17,18],tgt_file:10,tgt_file_path:13,tgt_group:3,tgt_lang:[3,20],tgt_languag:3,tgt_len:9,tgt_onmttok_kwarg:[15,17,18],tgt_pad_mask:11,tgt_prefix:13,tgt_sent:13,tgt_seq_length:[5,15,17,18],tgt_seq_length_trunc:[17,18],tgt_subword_alpha:[15,17,18],tgt_subword_model:[15,17,18],tgt_subword_nbest:[15,17,18],tgt_subword_typ:[5,15,17,18],tgt_subword_vocab:[15,17,18],tgt_vocab:[5,9,10,15,17,18],tgt_vocab_s:[5,17,18],tgt_vocab_threshold:[15,17,18],tgt_words_min_frequ:[17,18],than:[0,13,14,17,19],thang:21,thank:14,thant:13,thei:[2,5,13,14],them:[3,4,11],therefor:14,thi:[0,2,3,4,5,7,8,9,10,11,13,14,15,17,18,19,20],thin:9,thing:[0,3],thread:15,three:2,threshold:[15,17,18],through:[2,3,4,9,14,20],thu:9,tic:0,tick:0,tiedemann:8,time:[2,3,7,9,13,14,15,17,18,20],timeout:12,timer:12,timothe:8,titl:8,to_cpu:12,to_gpu:12,to_yyi:3,tok:12,token:[3,5,9,10,12,13,14,15,17,18],token_drop:6,token_mask:6,tokendrop:[15,17,18],tokendrop_temperatur:[15,17,18],tokenizer_mark:12,tokenizer_opt:12,tokenmask:[15,17,18],tokenmask_temperatur:[15,17,18],too:13,tool:[1,5,6,20],toolkit:[8,14],top:[0,2,13,18],topk_id:13,topk_scor:13,torch:[0,7,9,10,11,17],torchtext:9,total:[3,9,17],tqdm:19,trail:0,train:[1,3,6,7,8,9,10,11,14],train_:[4,5,20],train_ar:4,train_bg:[5,20],train_c:[5,20],train_d:5,train_da:5,train_el:5,train_en:[5,20],train_et:5,train_extremely_large_corpu:19,train_fi:5,train_fr:5,train_from:17,train_hu:5,train_it:[5,9],train_loss:9,train_loss_md:9,train_lt:5,train_lv:5,train_nl:5,train_pl:5,train_pt:5,train_ro:5,train_sk:5,train_sl:5,train_step:[9,17],train_sv:5,trainabl:9,trainer:6,training_step:9,transfer:14,transform:[1,2,5,6,9,10,11,14,21],transformer_decod:11,transformer_encod:11,transformer_ff:[5,17],transformerattentionbridgelay:6,transformerdecod:11,transformerdecoderbas:11,transformerencod:11,transformerencoderlay:2,transforms_cl:10,transit:11,translat:[1,3,4,5,6,8,9,11,12,14,16,21],translate_batch:13,translation_serv:12,translationbuild:13,translationserv:12,transvers:14,travi:0,trg:3,triangl:3,trichotomi:14,trick:[6,11],trivial:11,trunc_siz:9,truncat:[9,17,18],truncated_decod:17,trust:19,turn:[11,17],tutori:[4,6,7,20],two:[2,3,5,7,8,11,14],txt:[0,18,19],type:[0,2,3,6,8,9,10,11,12,13,15,18],typic:[9,17],typolog:14,under:[1,3,7,17,18],undergo:2,undergon:2,underli:13,understand:4,uniform:17,unigram:[15,17,18],union:0,uniqu:[4,6],unit:[2,14,19],unittest:0,univers:8,unk:[13,18],unknown:13,unless:3,unload:12,unload_model:12,unmodifi:13,unnecessari:[0,3],unpc:[4,6],unset:3,unshar:[],until:[13,18],unwieldi:3,updat:[7,9,12,13,17],update_finish:13,update_learning_r:17,update_n_src_word:9,update_vocab:17,upgrad:7,upon:1,upper:3,url:[7,8,21],url_root:16,usag:[5,6,15,16,17,18],use:[0,3,4,5,6,7,9,11,12,13,14,15,17,18,19,20],used:[2,3,5,9,10,11,12,13,14,15,17,18,19],useful:[1,9],user:[4,5,6,7,9,12],uses:[0,3,5,6,7,11,13,17],using:[0,1,2,3,4,5,8,11,12,13,15,17,18,19,20],usual:7,uszkoreit:21,util:[2,4,5,9,10,14],valid:[5,9,15,17,18],valid_batch_s:[5,17],valid_it:9,valid_loss:9,valid_loss_md:9,valid_step:[5,9,17],valu:[2,3,4,5,9,11,12,13,15,17,18,19],variabl:[3,4,7,10,13,20],variat:0,variou:[4,14],vaswani:21,vaswanispujgkp17:0,vector:[11,14,17],venv:7,verbos:[13,17,18],veri:[0,14,18],versatil:14,version:[1,5,12,13,19],via:[11,21],vinyal:21,virtual:7,visit:0,visual:[14,17,20],vocab:[4,5,6,9,13,14],vocab_path:[15,17,18],vocab_s:[13,17,19],vocab_sample_queue_s:15,vocab_size_multipl:[17,18],vocabs_dict:10,vocabulari:[3,4,5,9,14,15,17,18,19],vsp:[11,21],wai:[3,5,13],wait:[3,5],wang:21,want:[3,18],warmup:17,warmup_step:[5,17],warn:[10,15,17,18],websit:[5,19],weight:[2,3,11,17,18],weight_decai:17,weighted_sampl:17,well:[0,6,17],wget:[4,5,19,20],what:[3,6,9,12],when:[0,3,8,10,11,13,15,17,18,19],where:[2,3,6,7,11,13,14,15,17,18],wherea:[13,17],whether:[9,11,12,13,15,17,18],which:[1,3,4,11,13,14,17],whl:7,whole:13,whose:18,why:2,wiki:17,wikipedia:17,window:[15,17,18],with_align:9,within:[2,11,12],without:[0,7,17],wojciech:21,wolfgang:21,word2vec:[17,18],word:[2,11,13,14,15,17,18],word_align:13,word_lut:11,word_padding_idx:11,word_ratio_threshold:[15,17,18],word_vec_s:11,word_vocab_s:11,work:[0,3,7,13,17],workflow:8,workstat:20,world_siz:[5,17],would:[3,11,13,17],wpdn18:[15,17,18,21],wrap:[11,12],wrapper:[4,5,9],writabl:3,write:[3,5,9,19],writer:9,wsc:[13,21],www:17,xavier_uniform:17,xent:9,xinyi:21,xiong:21,xx_side_in:19,xx_side_out:19,xx_stream:19,xzvf:19,yaml:[4,5,15,17,18],year:[4,8,19],yet:13,yield:10,yml:[0,5,20],yonghui:21,yoon:8,you:[0,3,4,5,7,11,17,18,19,20,21],your:[0,3,4,6,7,19,20],your_config:20,your_path:5,your_project_name_is_your_account:7,your_venv_nam:7,your_vevn_nam:7,yourself:[8,19],yuan:21,yuntian:8,yyi:3,zaremba:21,zero:[3,9,11,13,15,17,18],zero_grad:9,zhang:21,zhifeng:21,ziemski:19,zihang:21,zip:19,zquez:8,zxs18:[11,21]},titles:["Contributors","About MAMMOTH","Attention Bridge","Config-config Tool","MAMMOTH Sharing Schemes","Training MAMMOTH 101","Contents","Installation","Overview","Framework","Data Loaders","Modules","Server","Translation","Component-level Modularity","Build Vocab","Server","Train","Translate","Prepare Data","Quickstart","References"],titleterms:{"case":1,"class":13,"function":1,"while":1,Are:1,The:3,about:1,actual:3,adapt:[3,17],adapter_config:3,ae_path:3,ae_transform:3,algorithm:1,align:17,allocate_devic:3,altern:3,anatomi:14,ani:1,approach:1,argument:16,attent:[2,11,17],autoencod:3,beam:18,between:1,bridg:[2,11,14,17],build:15,can:1,capabl:1,citat:8,cluster_languag:3,command:[3,4],common:[15,17,18],complete_language_pair:3,compon:14,config:3,config_al:3,config_config:3,configur:[5,15,17,18,20],content:6,contribut:1,contributor:0,core:12,corpora:3,corpora_schedul:3,creation:5,custom:14,data:[5,10,15,17,18,19,20],dataset:[4,10],dec_sharing_group:3,decod:[4,11,13,17,18],denois:[15,17,18],develop:1,direct:19,directori:5,distanc:3,distance_matrix:3,docstr:0,document:1,doe:1,domain:1,download:[5,19],dynam:17,effici:18,embed:[11,17,18],enc_sharing_group:3,encod:[4,11,17],enhanc:1,environ:5,europarl:19,featur:[1,17],feedforwardattentionbridgelay:2,filter:[15,17,18],flexibl:14,framework:9,fulli:4,further:20,futur:1,gener:17,get:19,group:3,guidelin:0,has:1,how:1,inferfeat:[15,17,18],initi:17,innov:1,input:3,instal:[7,8,20],introduc:1,issu:1,its:1,job:5,kei:[1,3],languag:3,level:[3,14],leverag:1,linattentionbridgelay:2,line:3,load:10,loader:10,log:[17,18],loss:9,lumi:7,made:1,mahti:7,mammoth:[1,4,5,20],manual:3,matrix:3,model:[5,9,12,14,17,18,19],modif:1,modul:11,modular:[4,14],n_gpus_per_nod:3,n_group:3,n_node:3,name:16,note:4,onmttok:[15,17,18],opennmt:1,optim:[9,17],opu:19,origin:1,other:3,overrid:3,overview:[4,8],own:1,parallel:14,paramet:[3,14],pars:19,particularli:1,path:19,penalti:18,perceiverattentionbridgelay:2,plan:1,prepar:[5,19,20],project:1,prune:[17,18],puhti:7,quickstart:20,random:18,rate:17,read:20,refer:21,relationship:1,relev:19,remove_temporary_kei:3,report:1,repositori:1,reproduc:[15,17,18],run:7,sampl:18,scheme:4,score:13,search:18,sentencepiec:[5,19],separ:4,server:[12,16],set:19,set_transform:3,setup:[4,5],share:[3,4,14],sharing_group:3,shot:19,simpleattentionbridgelay:2,specif:1,specifi:3,src_path:3,stage:3,start:20,statu:1,step:[5,19,20],strategi:13,structur:14,sublay:11,submiss:5,subword:[15,17,18],suit:1,supervis:19,switchout:[15,17,18],system:4,task:[5,15,17,18],test:[4,19],tgt_path:3,than:3,token:19,token_drop:[15,17,18],token_mask:[15,17,18],tool:3,top:3,train:[4,5,17,19,20],trainer:9,transform:[3,15,17,18],transformerattentionbridgelay:2,translat:[13,18,19,20],translation_config:3,translation_config_dir:3,trick:18,type:17,uniqu:1,unpc:19,unshar:4,usag:3,use:1,use_introduce_at_training_step:3,use_src_lang_token:3,use_weight:3,user:1,uses:1,valid:19,variabl:[5,19],vocab:[10,15,17,18,19],well:1,what:1,where:1,yaml:3,your:5,zero:19,zero_shot:3}}) \ No newline at end of file +Search.setIndex({docnames:["CONTRIBUTING","FAQ","attention_bridges","config_config","examples/sharing_schemes","examples/train_mammoth_101","index","install","main","mammoth","mammoth.inputters","mammoth.modules","mammoth.translate.translation_server","mammoth.translation","modular_model","options/build_vocab","options/server","options/train","options/translate","prepare_data","quickstart","ref"],envversion:{"sphinx.domains.c":1,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":1,"sphinx.domains.index":1,"sphinx.domains.javascript":1,"sphinx.domains.math":2,"sphinx.domains.python":1,"sphinx.domains.rst":1,"sphinx.domains.std":1,"sphinx.ext.viewcode":1,sphinx:56},filenames:["CONTRIBUTING.md","FAQ.md","attention_bridges.md","config_config.md","examples/sharing_schemes.md","examples/train_mammoth_101.md","index.rst","install.md","main.md","mammoth.rst","mammoth.inputters.rst","mammoth.modules.rst","mammoth.translate.translation_server.rst","mammoth.translation.rst","modular_model.md","options/build_vocab.rst","options/server.rst","options/train.rst","options/translate.rst","prepare_data.md","quickstart.md","ref.rst"],objects:{"mammoth.Trainer":{train:[9,1,1,""],validate:[9,1,1,""]},"mammoth.inputters.dataloader":{DynamicDatasetIter:[10,0,1,""],InferenceBatcher:[10,0,1,""],LookAheadBucketing:[10,0,1,""]},"mammoth.inputters.dataloader.DynamicDatasetIter":{from_opts:[10,1,1,""]},"mammoth.inputters.dataloader.LookAheadBucketing":{bucket_is_empty:[10,1,1,""],is_empty:[10,1,1,""],maybe_replenish:[10,1,1,""]},"mammoth.inputters.dataset":{ParallelCorpus:[10,0,1,""]},"mammoth.inputters.vocab":{Vocab:[10,0,1,""]},"mammoth.inputters.vocab.Vocab":{merge:[10,1,1,""]},"mammoth.models":{NMTModel:[9,0,1,""]},"mammoth.models.NMTModel":{count_parameters:[9,1,1,""],forward:[9,1,1,""]},"mammoth.modules":{Embeddings:[11,0,1,""]},"mammoth.modules.Embeddings":{emb_luts:[11,1,1,""],forward:[11,1,1,""],load_pretrained_vectors:[11,1,1,""],word_lut:[11,1,1,""]},"mammoth.modules.attention_bridge":{AttentionBridge:[11,0,1,""]},"mammoth.modules.attention_bridge.AttentionBridge":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.modules.average_attn":{AverageAttention:[11,0,1,""]},"mammoth.modules.average_attn.AverageAttention":{cumulative_average:[11,1,1,""],cumulative_average_mask:[11,1,1,""],forward:[11,1,1,""]},"mammoth.modules.decoder":{DecoderBase:[11,0,1,""]},"mammoth.modules.decoder.DecoderBase":{from_opts:[11,1,1,""]},"mammoth.modules.decoder_ensemble":{EnsembleModel:[11,0,1,""]},"mammoth.modules.encoder":{EncoderBase:[11,0,1,""]},"mammoth.modules.encoder.EncoderBase":{forward:[11,1,1,""]},"mammoth.modules.layer_stack_decoder":{LayerStackDecoder:[11,0,1,""]},"mammoth.modules.layer_stack_decoder.LayerStackDecoder":{add_adapter:[11,1,1,""],forward:[11,1,1,""],from_opts:[11,1,1,""],from_trans_opt:[11,1,1,""],get_submodule:[11,1,1,""],init_state:[11,1,1,""]},"mammoth.modules.layer_stack_encoder":{LayerStackEncoder:[11,0,1,""]},"mammoth.modules.layer_stack_encoder.LayerStackEncoder":{add_adapter:[11,1,1,""],forward:[11,1,1,""],from_opts:[11,1,1,""],from_trans_opt:[11,1,1,""],get_submodule:[11,1,1,""]},"mammoth.modules.mean_encoder":{MeanEncoder:[11,0,1,""]},"mammoth.modules.mean_encoder.MeanEncoder":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.modules.multi_headed_attn":{MultiHeadedAttention:[11,0,1,""]},"mammoth.modules.multi_headed_attn.MultiHeadedAttention":{forward:[11,1,1,""]},"mammoth.modules.position_ffn":{PositionwiseFeedForward:[11,0,1,""]},"mammoth.modules.position_ffn.PositionwiseFeedForward":{forward:[11,1,1,""]},"mammoth.modules.transformer_decoder":{TransformerDecoder:[11,0,1,""]},"mammoth.modules.transformer_decoder.TransformerDecoder":{forward:[11,1,1,""]},"mammoth.modules.transformer_encoder":{TransformerEncoder:[11,0,1,""]},"mammoth.modules.transformer_encoder.TransformerEncoder":{forward:[11,1,1,""],from_opts:[11,1,1,""]},"mammoth.translate":{BeamSearch:[13,0,1,""],DecodeStrategy:[13,0,1,""],GNMTGlobalScorer:[13,0,1,""],GreedySearch:[13,0,1,""],Translation:[13,0,1,""],TranslationBuilder:[13,0,1,""],Translator:[13,0,1,""]},"mammoth.translate.BeamSearch":{initialize:[13,1,1,""]},"mammoth.translate.DecodeStrategy":{advance:[13,1,1,""],block_ngram_repeats:[13,1,1,""],initialize:[13,1,1,""],maybe_update_forbidden_tokens:[13,1,1,""],maybe_update_target_prefix:[13,1,1,""],target_prefixing:[13,1,1,""],update_finished:[13,1,1,""]},"mammoth.translate.GreedySearch":{advance:[13,1,1,""],initialize:[13,1,1,""],update_finished:[13,1,1,""]},"mammoth.translate.Translation":{log:[13,1,1,""]},"mammoth.translate.Translator":{translate_batch:[13,1,1,""]},"mammoth.translate.greedy_search":{sample_with_temperature:[13,2,1,""]},"mammoth.translate.penalties":{PenaltyBuilder:[13,0,1,""]},"mammoth.translate.penalties.PenaltyBuilder":{coverage_none:[13,1,1,""],coverage_summary:[13,1,1,""],coverage_wu:[13,1,1,""],length_average:[13,1,1,""],length_none:[13,1,1,""],length_wu:[13,1,1,""]},"mammoth.translate.translation_server":{ServerModel:[12,0,1,""],ServerModelError:[12,3,1,""],Timer:[12,0,1,""],TranslationServer:[12,0,1,""]},"mammoth.translate.translation_server.ServerModel":{build_tokenizer:[12,1,1,""],detokenize:[12,1,1,""],do_timeout:[12,1,1,""],maybe_convert_align:[12,1,1,""],maybe_detokenize:[12,1,1,""],maybe_detokenize_with_align:[12,1,1,""],maybe_postprocess:[12,1,1,""],maybe_preprocess:[12,1,1,""],maybe_tokenize:[12,1,1,""],parse_opt:[12,1,1,""],postprocess:[12,1,1,""],preprocess:[12,1,1,""],rebuild_seg_packages:[12,1,1,""],to_gpu:[12,1,1,""],tokenize:[12,1,1,""],tokenizer_marker:[12,1,1,""]},"mammoth.translate.translation_server.TranslationServer":{clone_model:[12,1,1,""],list_models:[12,1,1,""],load_model:[12,1,1,""],preload_model:[12,1,1,""],run:[12,1,1,""],start:[12,1,1,""],unload_model:[12,1,1,""]},"mammoth.utils":{Optimizer:[9,0,1,""],Statistics:[9,0,1,""]},"mammoth.utils.Optimizer":{amp:[9,1,1,""],backward:[9,1,1,""],from_opts:[9,1,1,""],learning_rate:[9,1,1,""],step:[9,1,1,""],training_step:[9,1,1,""],zero_grad:[9,1,1,""]},"mammoth.utils.Statistics":{accuracy:[9,1,1,""],all_gather_stats:[9,1,1,""],all_gather_stats_list:[9,1,1,""],elapsed_time:[9,1,1,""],log_tensorboard:[9,1,1,""],output:[9,1,1,""],ppl:[9,1,1,""],update:[9,1,1,""],xent:[9,1,1,""]},"mammoth.utils.loss":{LossComputeBase:[9,0,1,""]},mammoth:{Trainer:[9,0,1,""]}},objnames:{"0":["py","class","Python class"],"1":["py","method","Python method"],"2":["py","function","Python function"],"3":["py","exception","Python exception"]},objtypes:{"0":"py:class","1":"py:method","2":"py:function","3":"py:exception"},terms:{"1node":5,"24th":8,"25g":7,"30gb":19,"3lang":19,"4gpu":5,"64k":[5,19],"666m":19,"abstract":11,"boolean":[9,13],"break":[14,19],"byte":17,"case":[4,6,20],"class":[0,6,9,10,11,12],"default":[12,15,16,17,18,19],"export":[4,5,7],"final":[2,5,11,13],"float":[3,11,13],"function":[0,2,3,4,6,9,11,12,13,17],"import":[0,19],"int":[9,10,11,12,13],"long":0,"new":[0,2,20],"portoro\u017e":19,"public":7,"return":[0,9,11,12,13],"short":14,"static":[9,17],"t\u00e4ttar":14,"throw":11,"true":[3,5,9,10,11,13,17,18,19],"try":[0,7,10],"while":[3,6,11,14],And:0,Are:6,But:7,EOS:13,For:[0,3,4,5,7,11,13,14,17,19,20],IDs:13,IFS:19,Its:1,LPs:3,Not:0,One:3,The:[2,4,5,7,9,11,12,13,14,17,19,20],Then:[0,5,19],There:[2,3],These:[2,3,4,13],Use:[3,17],Used:13,Will:3,__init__:12,_compute_loss:9,a3s:[4,5,19,20],aan:[11,17],aan_useffn:[11,17],ab_fixed_length:17,ab_lay:17,ab_layer_norm:17,abbrevi:0,abil:11,about:[0,6],abov:[0,11,13,20],abs:[2,17,18,21],acceler:[11,21],accept:[0,3,5,13],access:[2,3,7,20],accord:[3,4],accordingli:[],account:[3,7],accross:9,accum:[9,10],accum_count:[5,9,17],accum_step:[9,17],accumul:[9,17],accuraci:[5,9,13],achiev:[3,11,14],achin:21,acl:[8,21],aclantholog:[8,17],aclweb:17,across:[4,5,14],action:[11,13,17],activ:[1,2,7,11,17],activation_fn:11,activationfunct:11,actual:13,adadelta:17,adafactor:[5,17],adagrad:17,adagrad_accumulator_init:17,adam:17,adam_beta1:17,adam_beta2:17,adamoptim:17,adamw:17,adapt:[1,4,6,8,10,11,14],adapter_group:11,adapter_nam:3,add:[0,10,11],add_adapt:11,add_argu:[],added:3,adding:[0,1],addit:[0,4,11,15,17,18],addition:[4,11],address:[8,13],adjust:[3,4,20],adopt:[1,17],advanc:[13,17],advic:0,aforement:4,after:[0,2,4,13,17,18,19],afterward:11,again:0,against:11,aidan:21,alessandro:8,alexand:8,algorithm:[6,21],align:[6,9,11,12,13,18],align_debug:18,alignment_head:[11,17],alignment_lay:[11,17],aliv:13,alive_attn:13,alive_seq:13,all:[0,3,4,9,10,11,13,14,15,17,18,20,21],all_gather_stat:9,all_gather_stats_list:9,all_preprocess:12,allennlp:0,alloc:[3,5,14],allow:[0,1,2,3,4,14,17],almost:[13,17],alon:0,alpha:[13,18],alphabet:3,alreadi:[5,15,17,18,20],also:[0,3,4,5,7,9,11,14,17,19],altern:[4,11,19],although:11,alwai:[0,3,11],amd:7,among:[4,14],amp:[9,17],anatomi:6,ancestor:8,ancestri:14,ani:[0,3,6,13,15,17,18],anoth:[0,2,9],antholog:17,apex:17,apex_opt_level:17,api:[0,6],api_doc:17,appear:3,append:[7,19],appl:14,appli:[2,3,11,13,15,17,18],applic:18,approach:[6,14],appropri:[4,13],approxim:17,arab:[4,19],arbitrarili:20,architectur:[1,2,14],arg:[0,11,12],argmax:18,argpars:12,argument:[0,5,6,20],argumentpars:[],arn:8,around:19,arxiv:[0,2,17,18,21],ashish:21,ask:6,aspect:14,assign:[3,5,18],assum:[11,13,19,20],att_typ:2,attend:2,attent:[0,6,9,13,14,18,21],attention_bridg:[9,11],attention_dropout:[11,17],attentionbridg:11,attentionbridgenorm:2,attn:[11,13,18],attn_debug:[13,18],attn_typ:11,attr:12,attribut:[13,14],attributeerror:11,augment:21,author:8,autodoc:0,autogener:17,automat:[1,5,20],avail:[4,9,12,14,17,18,20],available_model:16,averag:[11,17,18,21],average_attn:11,average_decai:[9,17],average_everi:[9,17],average_output:11,averageattent:11,avg:[11,18],avg_raw_prob:18,avoid:[0,3],aws:7,axi:13,back:9,backend:17,backward:9,bahdanau:17,balanc:14,ban_unk_token:[13,18],bank:11,barri:21,bart:[5,15,17,18],base:[0,1,2,3,4,5,7,8,9,10,11,12,13,14,15,17,18,20],baselin:17,basemodel:9,basenam:19,bash:7,batch:[2,5,9,10,11,13,17,18,20],batch_siz:[4,5,10,11,13,17,18,20],batch_size_multipl:[10,17],batch_typ:[5,10,17,18,20],beam:[6,13],beam_search:13,beam_siz:[13,18],beamsearch:13,beamsearchbas:13,becaus:[3,18],becom:3,been:[13,15,17,18],befor:[0,4,12,13,17,18,19,20],begin:[9,13],behavior:14,below:[0,4,14,19,20],ben:3,bengali:3,best:[13,18],beta1:17,beta2:17,beta:[13,18],better:[0,4,7,15,17,18],between:[2,4,6,11,14,15,17,18,21],beyond:9,bia:11,biao:21,bib:0,bibtex:0,bibtext:0,bidir_edg:17,bidirect:17,bin:[7,17],binari:11,bit:18,bitext:14,blank:0,block:[13,18],block_ngram_repeat:[13,18],boggia:8,booktitl:8,bool:[9,10,11,12,13],bos:13,both:[3,4,13,14,17,18],both_embed:[17,18],bound:11,boundari:[15,17,18],box:20,bpe:[15,17,18],bptt:[9,17],bridg:[6,21],bridge_extra_nod:17,broad:14,broadcast:14,browser:0,bucket:10,bucket_fn:10,bucket_is_empti:10,buffer:9,build:[0,1,6,9,11,12,13,14,18,19,20],build_token:12,build_vocab:15,built:[0,9],bulgarian:5,bytetensor:13,cach:11,calcul:[2,9,13],call:[5,11,13],callabl:13,callback:9,can:[2,3,4,5,6,7,9,12,13,14,15,17,18,19,20],cancel:12,candid:[3,15,17,18],cao:21,capabl:[6,14],capit:0,captur:2,care:11,cat:19,catalan:14,categor:13,categori:13,celebr:5,central:14,challeng:[5,14,19],chang:[0,3,9,17],channel:2,charact:0,character:14,character_coverag:19,characterist:14,check:[0,5,8,10,11,20],checklist:0,checkpoint:[4,9,17,20],chen:21,chines:[4,19],chmod:7,choic:[0,10,11,15,17,18],choos:[0,1,4,5,15,17,18],chosen:[13,20],chronopoul:14,citat:[0,6],cite:[0,5,8,19],classmethod:[9,10,11],clean:[5,19],clear:0,clip:5,clone:12,clone_model:12,close:0,cls:9,cluster:[3,7,8,14,20],clutter:0,cmd:20,code:[0,3,7],code_dir:7,codebas:[5,7],coder:14,cohes:14,collect:10,column:3,com:[],combin:[14,18],comma:3,command:6,commenc:4,comment:0,common:[0,6,14],commun:[0,1,5,14],compar:8,compil:19,complet:[4,5,13,20],complex:[3,13,20],compon:[2,3,6,18],composit:17,comput:[3,4,8,9,11,14,17,18],concat:[11,17],conclus:14,condit:[13,17,18],conf:16,confer:8,config:[4,5,6,12,15,16,17,18,20],config_dir:5,config_fil:12,configur:[1,3,4,6,14],congratul:20,connect:2,consid:[1,3,14,19],consider:17,consist:[0,4,19],constant:3,constructor:[0,11],consum:17,contain:[3,4,5,7,11,12,13,14,19,20],content:[0,18],context:[2,11,17],context_attn:11,continu:0,contribut:[0,2,6],contributor:6,control:[3,9],conv2d:11,conv:11,conveni:[3,4],convent:0,convers:13,convert:[10,12],coordin:14,copi:[0,3,7,11,17,18,20],copy_attn:[11,13,17],copy_attn_forc:17,copy_attn_typ:17,copy_loss_by_seqlength:17,core:[2,6,9],corpora:[10,14],corpora_info:10,corpu:[3,5,10,14,15,17,18,19],corr:[0,21],correct:[3,4],correspand:12,correspond:[4,10,14,18],could:[13,14],count:[3,9,10,13,15,17,18],count_paramet:9,cov:13,cov_pen:13,coverag:[13,17,18],coverage_attn:17,coverage_non:13,coverage_penalti:[13,18],coverage_summari:13,coverage_wu:13,cpu:[10,12,17,18],crai:7,crayon:17,creat:[3,5,7,9,19,20],creation:3,criteria:[14,17],criterion:[5,9],critic:[17,18],cross:[9,11,17],csc:[14,19],csv:3,ct2_model:12,ct2_translate_batch_arg:12,ct2_translator_arg:12,ctrl:0,cuda_visible_devic:[5,20],cumbersom:3,cumul:[11,13,18],cumulative_averag:11,cumulative_average_mask:11,cur_dir:19,cur_len:13,current:[3,9,10,11,13,14,17],curricula:3,curriculum:3,custom:[4,6,12,17,19],custom_opt:12,customiz:4,cut:[0,19],cutoff:13,d_ff:11,d_model:11,dai:21,data:[1,2,3,4,6,9,13,14,21],data_path:19,data_typ:[9,10,13,17,18],databas:14,dataload:10,datapoint:14,dataset:[6,15,17,18,19,20],dataset_adapt:10,datasetadapt:10,datastructur:12,dblp:0,ddress:21,deal:[3,14],debug:[16,17,18],dec:[3,14],dec_lay:[5,17,20],dec_sharing_group:[4,5,20],decai:17,decay_method:[5,17],decay_step:17,decod:[2,3,6,9,14,20],decode_strategi:13,decoder_ensembl:11,decoder_typ:[5,17],decoderbas:[9,11],decodestrategi:13,deep:14,def:0,defin:[3,5,11,14,15,17,18,20],definit:[11,14,20],degre:11,delai:3,delet:[15,17,18],delimit:18,delv:4,deng:8,denois:[3,5,6],denoising_object:[15,17,18],denot:[2,4],depend:[0,3,7,9,12,19],deprec:[17,18],depth:14,desc:19,describ:[2,11,12,17,20],descript:0,design:[4,14,20],desir:[3,4,20],detail:[1,4,5,7,8,15,17,18,20],determin:[3,14,18],detoken:12,dev:[7,19],develop:[0,6],devic:[3,4,10,11,13,14,18],device_context:9,deyi:21,diagon:3,diagram:11,dict:[3,9,10,12,13,15,17,18],dict_kei:17,dictionari:[9,11,13,17],differ:[0,1,2,3,4,11,12,14,18,20],dim:11,dimens:[5,11,13,17],dimension:2,dir:19,direct:[0,3,8,13,14],directli:[0,5,18,19],directori:[3,4,7,12,17,20],disabl:17,discard:[15,17,18],discourag:17,disk:[17,18],displai:9,dist:9,distanc:[11,17],distinct:[4,14],distribut:[3,5,9,11,13,14,15,17,18,20],dive:4,divers:[15,17,18],divid:[2,3,17,18],divis:11,dmon:5,do_timeout:12,doc:0,document:[0,4,6,8,19,20],doe:[3,6,18],doesn:19,doi:8,doing:[3,18],domain:6,don:0,done:[5,13,14,19],dot:[2,11,17],dotprod:17,dowmunt:19,down:[13,14,15],download:[4,6,7,20],dozen:8,drop:19,dropout:[5,9,11,15,17,18],dropout_step:[9,17],due:17,dummi:[5,11],dump:[15,17,18],dump_beam:[13,18],dump_sampl:15,dump_transform:[17,18],duplic:20,dure:[11,12,14,17,18],dynam:[6,10,11,18],dynamicdatasetit:10,each:[3,4,5,11,13,14,15,17,18,19],earli:[5,17],earlier:[2,15,17,18],early_stop:[5,17],early_stopping_criteria:[5,17],earlystopp:9,eas:[3,14],easi:0,easili:3,echo:[5,19],ecosystem:1,edg:17,effect:[2,3,12,15],effici:[6,9,14,21],eight:[],either:[13,17],elaps:9,elapsed_tim:9,element:[2,3],els:19,emb_fil:11,emb_lut:11,embed:[2,6,14,15],embedding_s:11,embeddingless:17,embeddings_typ:[17,18],emerg:2,emploi:[2,4,9],empti:[5,10,11,13,15,17,18,19],en_side_in:19,en_side_out:19,en_stream:19,enabl:[1,5,11,14,17,18],enable_embeddingless:[11,17],enc:[3,14],enc_hidden:11,enc_lay:[5,17,20],enc_output:11,enc_sharing_group:[4,5,20],encod:[2,3,6,9,13,14,19,20],encoder_typ:[5,17],encoderbas:[9,11],encordec:[15,17,18],encount:[15,17,18],encout:[10,15,17,18],end:13,eng:3,english:[3,4,5,19],enhanc:[2,6,14],ensembl:18,ensemblemodel:11,ensur:[2,4,14],entir:[4,19],entri:0,entropi:9,env_dir:7,environ:[4,7,19,20],eos:13,epoch:17,epsilon:17,equal:[13,17],equat:11,equival:17,error:[0,11,15,17,18],especi:3,essenti:13,establish:2,etal:8,etc:5,eural:21,europarl:[5,6,20],europarl_data:[19,20],european:[5,19],evalu:[9,19],even:3,event:13,everi:[9,11,17,18],everyth:14,exactli:0,exampl:[0,3,4,5,7,10,11,14,15,17,18,19,20],exce:17,except:[0,12,15,17,18],exclusion_token:13,execut:[4,14,15,17,18],exist:[5,11,15,17,18,19],exist_ok:19,exp:17,exp_host:17,exp_id:5,expand:[5,19],expans:14,expect:[3,13,20],experi:[4,5,15,17,18],experiment:17,expert:1,explan:[],explicitli:14,explor:4,exponenti:17,extend:0,extern:0,extra:[7,11,17],extract:[5,19],facilit:[2,14],fail:13,fairseq:0,fals:[5,9,10,11,12,13,15,16,17,18,19],familiar:8,faro:8,faster:17,favorit:5,feat_0:18,feat_1:18,feat_dim_expon:11,feat_merg:[11,17],feat_merge_s:17,feat_padding_idx:11,feat_vec_expon:[11,17],feat_vec_s:[11,17],feat_vocab_s:11,feats0:18,feats1:18,featur:[2,6,9,11,14,15,18,21],fed:2,feed:[3,11,14,17],feedforward:[2,17],feedforwardattentionbridgelay:6,feel:[0,4],few:[0,20],ffn:[11,17],field:10,fifth:5,figur:[11,14],file:[0,3,4,5,10,12,15,17,18,19,20],filenam:17,filter:[5,6,19],filterfeat:[15,17,18],filternonzeronumer:[15,17,18],filterrepetit:[15,17,18],filterterminalpunct:[15,17,18],filtertoolong:[3,5,15,17,18],filterwordratio:[15,17,18],find:[0,5],finer:14,firefox:0,first:[0,3,5,7,11,13,17],five:2,fix:[0,13,17],flag:[3,4,9],flake8:0,flexibl:[1,4],floattensor:[9,11,13],flow:[2,14],fnn:11,focu:[0,2],focus:14,folder:0,follow:[0,2,3,4,18,19,20],foo:0,forbidden:13,forbidden_token:13,forc:[13,18],format:[0,12,15,17,18,19,20],former:11,forward:[3,9,11,14,17],fotranmt:8,found:[19,20],foundat:[2,4],four:[],fp16:[17,18],fp32:[9,17,18],frac:2,fraction:[15,17,18],framework:[4,6,17],free:[0,4,5,12,19],freez:[11,17],freeze_word_vec:11,freeze_word_vecs_dec:17,freeze_word_vecs_enc:17,french:[4,19],frequenc:[15,17,18],frequent:6,friendli:1,from:[1,2,3,5,7,9,10,11,13,14,17,18,19,20],from_opt:[9,10,11],from_trans_opt:11,from_xxx:3,frozenset:13,full:[0,3,11,12,14,15,17,18,19],full_context_align:[11,17],fulli:[1,3,11,14],further:[5,15,17,18,19],fusedadam:17,futur:[6,7],gao:21,gap:21,garg:17,gather:9,gating_output:11,gelu:17,gener:[0,1,2,3,4,5,6,9,10,13,18,19,20],generator_funct:17,ger:14,get:[6,7],get_submodul:11,git:[],github:[17,19],give:[3,17,18],given:[2,3,11,12],global_attent:17,global_attention_funct:17,global_scor:13,glove:[17,18],gnmt:13,gnmtglobalscor:13,going:13,gold:13,gold_scor:13,gold_sent:13,gomez:21,gone:17,good:[0,17],googl:[0,13,18,21],gpu:[3,4,5,7,12,13,14,17,18,20],gpu_backend:17,gpu_load:5,gpu_rank:[5,17,20],gpu_verbose_level:[9,17],gpuid:17,grad:9,gradient:[5,9,14,17],graham:21,grain:14,gram:13,graph:17,gre:7,greater:13,greedy_search:13,greedysearch:13,group:[4,5,11,14,17,18],groupwis:[3,14],grow:13,gtx1080:18,guid:[4,5,8,11,20],guidelin:[6,19],guillaum:8,hack:14,had:18,haddow:21,hand:[3,14],handl:[0,9,14],happen:13,hardwar:14,has:[2,3,4,6,11,13,14,15,17,18],has_cov_pen:13,has_len_pen:13,has_tgt:13,have:[0,3,11,13,14,17,18,19,20],head:[2,5,11,17],head_count:11,help:[0,2,18],helsinki:8,henc:20,here:[4,5,7,13,19],hidden:[9,11,17],hidden_ab_s:17,hidden_dim:2,hieu:21,high:[3,5,19],higher:[13,17,18],highest:18,hold:13,hook:11,hoorai:5,host:7,hour:5,how:[0,6,11,20],howev:[0,9],html:[0,17],http:[2,4,5,7,8,17,18,19,20,21],huge:17,human:[3,21],hyp:[4,20],hyper:1,hyphen:3,identifi:[5,18],idl:3,ids:3,ignor:[5,11,15,17,18],ignore_when_block:[13,18],iii:14,illia:21,illustr:14,ilya:21,imag:9,impact:17,implement:[2,9,14,17],improv:[1,5,11,13,17,19,21],in_config:3,in_featur:11,includ:[0,1,3,4,5,11,14,15,17,18,19],incompat:[15,17,18],incorpor:17,increas:[3,14],independ:14,index:[7,11,17],indic:[4,5,9,11,13,15,17,18],individu:[3,11],inf:13,infer:[10,13,14],inferencebatch:10,inferfeat:6,info:[5,10,17,18],inform:[2,3,5,14,17,18,20],infrastructur:14,ingredi:13,init:17,init_st:[9,11],initi:[4,5,6,9,11,12,13],initial_accumulator_valu:17,initil:10,inner:11,innov:6,inp:13,inp_seq_len:13,inproceed:8,input:[2,6,9,10,11,12,13,15,17,18,19,20,21],input_dir:19,input_len:11,input_sentence_s:19,inputs_len:11,inputt:[10,13],insert:[15,17,18],insert_ratio:[15,17,18],insid:5,instal:[0,5,6,19],instanc:[9,11,13],instanti:9,instead:[0,3,7,11,15,17,18],instruct:[7,17,19],int8:18,integ:13,integr:0,interact:7,interfac:[9,11],intermedi:2,intermediate_output:2,intern:12,interv:17,introduc:[2,3,6],introduct:3,invalid:[11,15,17,18],involv:[2,14],is_empti:10,is_finish:13,is_normform:11,is_on_top:11,is_train:10,island:8,isn:13,issu:6,item:[10,11],iter:[9,10],iterabledataset:10,its:[0,3,6,14],itself:[3,11],jakob:21,jean:8,jinsong:21,job:[4,7],johnson:14,joiner:[15,17,18],jone:21,journal:0,json:16,junczi:19,just:19,kaiser:21,keep:[12,13,17],keep_checkpoint:[5,17],keep_stat:17,keep_topk:13,keep_topp:13,kei:[6,11,14],kera:17,kernel_s:11,key_len:11,kim:8,klau:21,klein:8,koehn:5,krikun:21,kwarg:11,label:[5,17],label_smooth:[5,17],lakew:14,lambda:[15,17,18],lambda_align:17,lambda_coverag:17,lang:[3,5,19],lang_a:3,lang_b:3,lang_pair:[3,20],languag:[2,4,5,14,15,17,18,19,20],language_pair:19,larg:14,last:[3,4,17,18,19],latter:11,launch:[],layer:[2,5,11,14,17,18,20],layer_cach:11,layer_norm_modul:11,layer_stack_decod:11,layer_stack_encod:11,layer_stack_index:11,layer_type_to_cl:2,layernorm:17,layerstack:3,layerstackdecod:11,layerstackencod:11,layerwis:14,lead:13,learn:[2,5,9,14,17],learning_r:[5,9,17],learning_rate_decai:17,learning_rate_decay_fn:9,least:0,leav:[3,17,19],len:[9,11,13],length:[3,9,11,13,15,17,18,19],length_averag:13,length_non:13,length_pen:13,length_penalti:[13,18],length_wu:13,less:3,let:[3,4,11,14],level:[5,6,10,15,17,18,19],leverag:[6,14],lib:7,librari:[8,17],lightweight:14,like:[0,4,11,13,14,18,19],limit:18,lin:[2,14,17],linattentionbridgelay:6,line:[0,5,10,15,17,18,19],linear:[2,11],linear_warmup:17,linguist:[8,11,14,21],link:[0,2,7],list:[0,3,9,11,12,13,15,17,18],list_model:12,literatur:17,llion:21,load:[6,7,9,11,12,14,17,18],load_model:12,load_pretrained_vector:11,loader:6,local:[0,3],localhost:17,locat:5,log:[5,6,9,13],log_dir:[5,20],log_fil:[17,18],log_file_level:[17,18],log_prob:13,log_tensorboard:9,logger:13,login:7,logit:[13,18],logsumexp:13,longer:18,longest:13,longitudin:14,longtensor:[9,11,13],look:[0,8,10,11,18,19],look_ahead_s:10,lookaheadbucket:10,loop:9,loppi:8,loss:[6,17],loss_scal:17,losscomputebas:9,love:[0,1],low:14,lower:[3,17,18],lrec:19,lsl:[13,21],lstm:17,lua:12,lukasz:21,lumi:6,luong:[17,21],lust:7,lustrep1:7,lustrep2:7,macherei:21,machin:[1,5,8,11,13,14,17,18,20,21],made:[3,5,6],magic:13,mahti:6,mai:[3,5,8,9,10,12,13,15,17,19],main:[0,5,8,9,15,17,18,19],maintain:[4,13],make:[0,4,7,9,10,14,15,17,18,19],make_shard_st:9,mammoth101:[5,19,20],mammoth:[0,3,6,7,8,9,10,11,12,13,14,17,19],mammoth_transform:[15,17,18],mamooth:[19,20],manag:[4,5,9,14],mani:[9,13,17],manipul:9,manual:[4,5,12,13,19],map:[3,9,19],marian:17,mark:17,marker:12,mask:[11,15,17,18],mask_length:[15,17,18],mask_or_step:11,mask_ratio:[5,15,17,18],mass:[15,17,18],massiv:[1,3,8],master:[5,17],master_ip:[4,17],master_port:[4,5,17],match:12,mathbb:2,mathbf:2,matric:2,matrix:[2,11,17],max:[9,11,13,19],max_generator_batch:[5,17],max_grad_norm:[5,9,17],max_length:[13,18],max_relative_posit:[11,17],max_sent_length:18,max_sentence_length:19,max_siz:9,maxim:21,maximum:[15,17,18],maybe_convert_align:12,maybe_detoken:12,maybe_detokenize_with_align:12,maybe_postprocess:12,maybe_preprocess:12,maybe_replenish:10,maybe_token:12,maybe_update_forbidden_token:13,maybe_update_target_prefix:13,mean:[3,4,11,12,14,17,18],mean_encod:11,meanencod:11,mechan:[1,2,3,14],mem:7,memori:[11,12,17,19],memory_bank:[11,13],memory_length:11,merg:[10,11,17],meta:3,metadata:9,method:[9,11,17],metric:18,mi250:7,michel:8,micku:8,mike:21,million:8,min_length:[13,18],minh:21,minim:14,minimum:[15,17,18],mirror:17,mix:9,mixer:10,mixingstrategi:10,mixtur:1,mkdir:[5,7,19],mlp:[11,17],mnmt:14,mode:[3,15,17,18],model:[1,2,3,4,6,11,13,15,20],model_checkpoint:20,model_dim:[5,11,17,20],model_dtyp:[9,17],model_fil:19,model_id:12,model_kwarg:12,model_prefix:19,model_root:12,model_sav:9,model_task:17,model_typ:[5,17],modelsaverbas:9,modif:[6,9],modifi:[0,4,13],modul:[0,2,6,7,9,14,17,20],modular:[1,6,8,20],module_id:11,mohammad:21,moment:5,monitor:5,monolingu:3,month:8,more:[0,1,3,5,7,10,13,14,15,17,18,20],most:[13,14,18],mostli:9,move:[12,17],moving_averag:[9,17],much:17,mul:[5,19,20],multi:[0,2,5,11,20],multi_headed_attn:11,multiheadedattent:[2,11],multilingu:[1,3,5,8,19,20],multinod:5,multipl:[0,2,3,9,10,11,14,17,18],multipli:[2,10],multplic:0,must:[3,11,12,14,17],my_config:20,mymodul:7,n_batch:9,n_best:[12,13,18],n_bucket:[10,17],n_correct:9,n_edge_typ:17,n_node:[5,17],n_sampl:[15,17,18],n_seg:12,n_src_word:9,n_step:17,n_word:9,naacl:17,name:[0,3,5,6,11,13,15,17,18,19,20],named_modul:11,namespac:12,napoleon:0,narg:[],nation:19,nccl:17,necessari:[0,4,5,7,9,13,17,18,20],necessit:3,need:[0,3,4,5,9,11,14,17,19,20,21],neg:[12,17],nest:11,net_b:11,net_c:11,network:[1,11,21],neubig:21,neural:[1,8,11,13,14,21],never:13,next:[3,9,13,18],nfeat:11,ngram:[13,18],nightmar:3,niki:[8,21],nlp:[8,20],nmt:[9,13,17,18],nmtmodel:[9,11],noam:[17,21],noamwd:17,nodalida:8,node:[3,5,7,9,14,17,20],node_gpu:[5,20],node_rank:[4,5,17],nois:3,non:[11,13,17,20],none:[5,9,10,11,12,13,15,17,18],nonetyp:13,nonzero_threshold:[15,17,18],nordic:8,norm:[11,17],norm_method:9,normal:[2,3,5,9,11,17],normalz:9,normform:[11,17],norouzi:21,note:[0,3,7,13,19,20],noth:[0,9],notset:[17,18],now:[1,4,5,20],nroo:8,ntask:7,nucleu:18,num_lay:11,num_step:9,num_thread:15,number:[3,9,10,11,13,14,15,17,18],numel_fn:10,numer:[15,17,18],nvidia:[5,7,17],obj:[0,9],object:[0,9,10,12,13,15,17,18,19],occur:4,oder:3,off:17,offer:[4,14],offici:[4,19],offset:10,ofi:7,often:[15,17,18],on_timemout:12,on_timeout:12,onc:[13,17],one:[0,2,3,4,5,9,10,11,15,17,18,20],ones:20,onli:[3,4,9,10,13,14,15,17,18,19],onmt:[5,19,20],onmt_token:[15,17,18],onmttok:6,open:[1,8,19],opennmt:[0,3,6,8,9,16],oper:14,operatornam:2,opt:[5,9,10,11,12,17,18],opt_level:17,optim:[1,5,6,14],option:[0,3,4,5,7,9,10,11,12,13,15,17,18,19,20],opu:[5,6],opus100:19,opustc:[5,19,20],ord:21,order:[3,17],org:[2,7,8,17,18,21],organ:[10,14],origin:[5,6,17,19],oriol:21,other:[2,4,5,7,9,13,15,17,18,19,20,21],other_lang:19,otherwis:[3,11,17,18],our:[5,7,13,19,20],our_stat:9,out:[3,8,9,20],out_config:3,out_featur:11,out_fil:13,out_path:[4,20],out_typ:19,output:[2,3,4,5,9,11,12,13,15,17,18,19,20],output_dir:[19,20],output_model:18,over:[0,3,9,13,17,18,19,20],overal:2,overcom:14,overhead:14,overrid:[11,13,15,17,18,20],overridden:11,overse:14,overview:6,overwrit:[5,7,15,17,18],own:[6,9,14],ownership:9,p17:8,p18:17,packag:[5,7,12],pad:[9,11,13],page:[5,8],pair:[3,4,9,12,14,17,19,20],paper:[0,2,5,17],parallel:[4,5,6,11,13,15,17,19],parallel_path:13,parallelcorpu:[10,13],param:9,param_init:[5,17],param_init_glorot:[5,17],paramet:[1,4,5,6,8,9,10,11,12,13,15,17,18,20],parenthes:0,parliament:[5,19],parmar:21,pars:[10,12],parse_arg:[],parse_opt:12,parser:[],part:[2,13],partial:14,particular:[0,3],particularli:6,partit:7,pass:[2,3,9,11,12,17],past:[0,17,20],path:[3,4,5,7,10,11,12,13,15,17,18,20],path_src:[5,20],path_tgt:[5,20],path_to_codebas:5,path_to_europarl:5,path_to_mammoth:[],path_to_src_languag:20,path_to_vocab:5,path_to_your_env:5,path_valid_src:5,path_valid_tgt:5,pathlib:19,patienc:[5,9],pattern:[3,15,17,18],pdf:17,pen:13,penalti:[6,13,15,17],penaltybuild:13,peopl:7,per:[0,3,15,17,18],perceiv:[2,17],perceiverattentionbridgelay:6,percentag:[15,17,18],perfom:17,perform:[2,11,17],permut:[15,17,18],permute_sent_ratio:[15,17,18],perplex:9,pertain:14,pfs:7,pham:21,philipp:5,phrase_t:[13,18],phylogenet:14,pip3:7,pip:[0,7,8,19,20],pipelin:[15,17,18],plain:10,plan:6,platform:7,pleas:[0,5,8,19,20],plu:17,point:[5,20],pointer:5,poisson:[15,17,18],poisson_lambda:[15,17,18],polosukhin:21,pool:[11,17],pool_siz:[10,17],port:[5,16,17],portal:8,pos_ffn_activation_fn:[11,17],posit:[5,11,17],position_encod:[5,11,17],position_ffn:11,positionalencod:11,positionwisefeedforward:[11,17],possibl:[3,9,11,12,13,15,17,18,20],postprocess:12,postprocess_opt:12,potenti:13,pouliquen:19,pouta:19,ppid:5,ppl:9,pre:[9,12,13],pre_word_vecs_dec:17,pre_word_vecs_enc:17,preced:3,precis:9,pred:18,pred_scor:13,pred_sent:13,predict:[9,13,18],prefer:0,prefix:[3,9,15,17,18],prefix_seq_len:13,preload:12,preload_model:12,prepar:[6,13],preprint:21,preprocess:[4,12,14,19],preprocess_opt:12,presenc:3,present:5,presum:13,pretrain:[11,17],prevent:[13,18],previou:[3,11,13,20],previous:2,primari:3,prime:2,print:[9,17,18,19],prior_token:[15,17,18],priori:14,prob:13,proba:18,probabl:[11,13,15,17,18],problem:[13,14],proc:[8,21],proce:20,procedur:3,proceed:[5,8,19],process:[2,4,5,9,12,14,15,17,19],processed_data:4,processu:12,produc:[2,13,15,17,18,19],product:2,progress:5,projappl:7,project:[0,2,5,6,7,8],project_462000125:7,propag:9,proper:12,properli:7,properti:[9,11],proport:[3,15,17,18],provid:[4,5,7,8,18,19,20],prune:6,pty:7,publish:8,puhti:6,pull_request_chk:0,punct_threshold:[15,17,18],punctuat:[0,15,17,18],purason:14,put:[13,20],pwd:19,pyonmttok:[15,17,18],python3:[3,4,7,20],python:[0,3,4,5,7,17,20],pythonpath:7,pythonuserbas:[5,7],pytorch:[0,1,7],qin:21,qualifi:11,quantiz:18,queri:11,query_len:11,question:6,queue:[15,17],queue_siz:17,quickstart:[6,8],quoc:21,quot:0,raganato:8,rais:[11,15,17,18],random:[6,15,17,19],random_ratio:[15,17,18],random_sampling_temp:[13,18],random_sampling_topk:[13,18],random_sampling_topp:[13,18],randomli:[5,13,19],rang:18,rank:[5,13,14,17],ranslat:21,rare:13,rate:[5,6,9],rather:0,ratio:[13,15,17,18],raw:[10,13,18],raw_prob:11,rccl:7,reach:13,read:[0,3,5,12,19],readabl:[0,3,17,18],readm:17,real:11,rebuild:12,rebuild_seg_packag:12,receiv:3,recent:17,recip:11,recommend:[4,7,17,20],recommonmark:0,rectifi:2,recurr:11,reduc:14,redund:3,ref:0,refer:[0,1,2,4,6,7,11,19,20],referenc:11,regardless:[3,4],regist:11,regular:[15,17,18],rel:[11,14,17],relat:[5,14,15,17,18],relationship:[2,6],releas:[5,7,19],relev:[11,13,14],relu:[2,11,17],rememb:[0,4],remov:[3,5,15,17,18,19],renorm:17,reorder:13,rep_max_len:[15,17,18],rep_min_len:[15,17,18],rep_threshold:[15,17,18],repeat:[13,15,17,18],repetit:18,replac:[4,13,15,17,18],replace_length:[5,15,17,18],replace_unk:[13,18],replic:11,report:[6,8,9,17,18],report_align:[13,18],report_everi:[5,17],report_manag:9,report_scor:13,report_stats_from_paramet:[9,17],report_tim:[13,18],reportmgrbas:9,repositori:6,repres:[5,9],represent:[2,11,14,17],reproduc:6,requir:[0,4,9,11,14,17,19,20],research:[5,8],reservoir:10,reset:9,reset_optim:17,resett:17,residu:[11,14],resolv:11,resourc:[3,5,14,19],respect:[2,3],respons:9,rest:16,restrict:[15,17,18],result:[11,12,17],return_attent:13,reus:17,reuse_copy_attn:17,revers:[15,17,18],reversible_token:[15,17,18],rico:21,right:0,rmsnorm:17,rnn:[9,17],roblem:21,rocm5:7,rocm:7,root:[2,3],rotat:[15,17,18],rotate_ratio:[15,17,18],roundrobin:17,row:3,rshavn:8,rsqrt:17,rst:0,run:[0,3,4,5,9,11,12,17,18,19,20],runtim:11,rush:8,russian:[4,19],s_idx:10,sacrebleu:7,sai:[3,11,14],samantao:7,same:[0,3,11,12,14,17,20],sampl:[6,13,15,17,19],sample_with_temperatur:13,sampling_temp:13,save:[5,9,14,15,17,18,19,20],save_all_gpu:[5,17],save_checkpoint_step:[5,9,17],save_config:[15,17,18],save_data:[15,17,18],save_dir:5,save_model:[5,17,20],saver:9,scalabl:[1,14],scale:[11,13,14,17],scenario:[1,20],schedul:[4,9,17],scheme:[5,6,14,20],schuster:21,score:[6,12,15,17,18],scorer:13,scratch:[1,7],script:[0,4,5,6,7,19,20],script_dir:5,seamless:4,search:[0,3,6,13],second:[2,3,11,12],section:[5,7],secur:[10,15,17,18],see:[3,5,11,12,13,15,17,18],seed:[5,13,15,17,18],seemingli:17,seen:2,segment:[3,12,15,17,18],select:[11,13],select_index:13,selector:3,self:[2,11,12,13,17],self_attn_typ:[11,17],send:[0,17],senellart:8,sennrich:21,sensibl:0,sent:[9,10,17,18],sent_numb:13,sentenc:[13,15,17,18,19],sentencepiec:[3,6,7,15,17,18,20],sentencepieceprocessor:19,separ:[3,5,11,14,19],seper:12,seq2seq:[13,17],seq:13,seq_len:[2,13],sequenc:[1,2,3,9,11,12,13,14,15,17,18],seri:14,serial:11,serv:[2,4],server:[6,17,19],servermodel:12,servermodelerror:12,session:7,set:[1,2,3,4,5,7,9,11,12,13,14,15,17,18,20],sever:[3,5,11,13],sgd:17,sh16:[11,21],shaham:17,shallow:14,shape:[0,11,13],shard:[9,17,18],shard_siz:[4,9,18,20],share:[1,5,6,7,8,11,15,17,18,19,20],share_decoder_embed:17,share_embed:17,share_vocab:[15,17,18],shared_dec:20,shared_enc:20,shazeer:21,shortest:13,shot:3,should:[3,5,11,13,17,18,20],show:[11,20],shuf:19,shuffl:[5,19],shuffle_input_sent:19,side:[3,9,12,15,17,18],side_a:3,side_b:3,sign:[15,17,18],signifi:4,silent:[5,11,15,17,18],similar:[2,3,11,17],simpl:[2,9,11,17],simpleattentionbridgelay:6,simpli:[11,20],simulatan:11,sin:17,sinc:11,singl:[0,4,5,12,14,17,20],single_pass:17,singular:7,site:7,six:[4,19],size:[3,5,9,10,11,13,15,17,18,19,20],skip:[3,5,15,17,18,19],skip_embed:11,skip_empty_level:[5,10,15,17,18],slovenia:19,slow:[15,18],slurm:[3,4,5,7,20],slurm_nodeid:[4,5],slurmd_nodenam:4,smaller:[14,15,17,18],smi:5,smooth:[5,15,17,18],softmax:[2,17,18],solid:4,some:[0,3,9,11,18],someth:[0,11],sometim:0,soon:5,sort:[12,19],sorted_pair:3,sourc:[0,1,3,4,5,7,8,9,10,11,12,13,14,15,17,18,20],sp_path:19,space:[0,2,14,17],spacer:[15,17,18],span:[4,15,17,18],spanish:[4,19],spars:[1,11],special:10,specif:[2,3,4,5,6,8,13,14,15,17,18,20],specifi:[4,5,11,15,17,18,20],sphinx:0,sphinx_rtd_them:0,sphinxcontrib:0,spill:0,spm:[5,19],spm_encod:19,spm_train:19,sqrt:2,squar:[2,3],src:[3,4,9,11,12,13,15,17,18,19,20],src_embed:[17,18],src_feat:18,src_feats_vocab:[15,17,18],src_file:10,src_file_path:13,src_ggnn_size:17,src_group:3,src_lang:[3,20],src_languag:3,src_len:[9,11],src_length:13,src_map:13,src_onmttok_kwarg:[15,17,18],src_raw:13,src_seq_length:[5,15,17,18],src_seq_length_trunc:[17,18],src_subword_alpha:[15,17,18],src_subword_model:[15,17,18],src_subword_nbest:[15,17,18],src_subword_typ:[5,15,17,18],src_subword_vocab:[15,17,18],src_tgt:[5,20],src_vocab:[5,10,13,15,17,18,20],src_vocab_s:[5,17,18],src_vocab_threshold:[15,17,18],src_words_min_frequ:[17,18],srun:[4,5,7],stabl:2,stack:[11,14,17],stage:2,stand:0,standalon:4,standard:[11,17,18],start:[3,4,5,6,7,9,12,17,19],start_decay_step:17,stat:[9,17],stat_list:9,state:[9,11,13,14,17],state_dict:17,state_dim:17,statist:[5,9,17],statu:6,stdout:9,step:[2,3,6,9,11,13,17,18],stepwis:11,stepwise_penalti:[13,18],stig:8,still:0,stop:[5,15,17,18],store:17,str:[0,9,10,11,12,13,19],strategi:[1,6,9,10,14,17],streamlin:14,stride:[10,11],string:[9,11,15,17,18],strip:19,structur:[2,4,6,17,18],structured_log_fil:[17,18],style:[0,10,11,14,15,17,18],styleguid:0,sub_id:11,subclass:[9,11,13],subcompon:3,subdirectori:7,sublay:6,submodul:11,subsequ:2,subset:19,substitut:3,substr:[15,17,18],subword:[3,6,14],successfulli:20,suggest:17,suit:[4,6],sum:[9,11,13,17],sume:9,summari:[0,5,13,18,19],summit:5,superclass:0,supercomput:7,supervis:[3,11,17],support:[0,1,3,17],suppos:19,sure:[4,7,13,19],sutskev:21,swahili:14,switchout:[6,21],switchout_temperatur:[15,17,18],symmetr:3,sync:14,synchron:14,system:[6,13,14,17,20,21],t_idx:10,tab:[15,17,18],tabl:[11,18],tag:10,tailor:4,take:[2,3,5,8,11,15,17,18,19],tar:[4,5,19,20],tarbal:19,target:[3,4,5,9,11,12,13,14,15,17,18],target_prefix:13,tartu:8,task2gpu:14,task:[3,6,9,10,11,13,14,20],task_distribution_strategi:17,task_id:[4,18,20],task_queue_manag:[9,10,11],tatoeba:[3,5,19],tau:[15,17,18],team:7,technic:8,temperatur:[3,13,15,17,18],templat:3,tensor:[0,9,10,11,13],tensorboard:[5,9,17,20],tensorboard_log_dir:[5,17,20],tensorflow:17,termin:[15,17,18],test:[0,1,4,7],text:[5,9,10,13,17,18,19,20],tgt:[3,9,11,12,15,17,18],tgt_embed:[17,18],tgt_file:10,tgt_file_path:13,tgt_group:3,tgt_lang:[3,20],tgt_languag:3,tgt_len:9,tgt_onmttok_kwarg:[15,17,18],tgt_pad_mask:11,tgt_prefix:13,tgt_sent:13,tgt_seq_length:[5,15,17,18],tgt_seq_length_trunc:[17,18],tgt_subword_alpha:[15,17,18],tgt_subword_model:[15,17,18],tgt_subword_nbest:[15,17,18],tgt_subword_typ:[5,15,17,18],tgt_subword_vocab:[15,17,18],tgt_vocab:[5,9,10,15,17,18,20],tgt_vocab_s:[5,17,18],tgt_vocab_threshold:[15,17,18],tgt_words_min_frequ:[17,18],than:[0,13,14,17,19],thang:21,thank:14,thant:13,thei:[2,5,13,14],them:[3,4,11],therefor:14,thi:[0,2,3,4,5,7,8,9,10,11,13,14,15,17,18,19,20],thin:9,thing:[0,3],thread:15,three:[2,19],threshold:[15,17,18],through:[2,3,4,9,14],thu:[9,20],tic:0,tick:0,tiedemann:8,time:[2,3,7,9,13,14,15,17,18,20],timeout:12,timer:12,timothe:8,titl:8,to_cpu:12,to_gpu:12,to_yyi:3,tok:12,token:[3,5,9,10,12,13,14,15,17,18,20],token_drop:6,token_mask:6,tokendrop:[15,17,18],tokendrop_temperatur:[15,17,18],tokenizer_mark:12,tokenizer_opt:12,tokenmask:[15,17,18],tokenmask_temperatur:[15,17,18],too:13,tool:[1,5,6,20],toolkit:[8,14],top:[0,2,13,18],topk_id:13,topk_scor:13,torch:[0,7,9,10,11,17],torchtext:9,total:[3,9,17],tqdm:19,trail:0,train:[1,3,6,7,8,9,10,11,14],train_:[4,5,20],train_ar:4,train_bg:[5,20],train_c:[5,20],train_d:5,train_da:5,train_el:5,train_en:[5,20],train_et:5,train_extremely_large_corpu:19,train_fi:5,train_fr:5,train_from:17,train_hu:5,train_it:[5,9],train_loss:9,train_loss_md:9,train_lt:5,train_lv:5,train_nl:5,train_pl:5,train_pt:5,train_ro:5,train_sk:5,train_sl:5,train_step:[9,17],train_sv:5,trainabl:9,trainer:6,training_step:9,transfer:14,transform:[1,2,5,6,9,10,11,14,20,21],transformer_decod:11,transformer_encod:11,transformer_ff:[5,17],transformerattentionbridgelay:6,transformerdecod:11,transformerdecoderbas:11,transformerencod:11,transformerencoderlay:2,transforms_cl:10,transit:11,translat:[1,3,4,5,6,8,9,11,12,14,16,21],translate_batch:13,translation_serv:12,translationbuild:13,translationserv:12,transvers:14,travi:0,trg:3,triangl:3,trichotomi:14,trick:[6,11],trivial:11,trunc_siz:9,truncat:[9,17,18],truncated_decod:17,trust:19,turn:[11,17],tutori:[4,6,7,19,20],two:[2,3,5,7,8,11,14,20],txt:[0,18,19],type:[0,2,3,6,8,9,10,11,12,13,15,18],typic:[9,17],typolog:14,under:[1,3,7,17,18],undergo:2,undergon:2,underli:13,understand:4,uniform:17,unigram:[15,17,18],union:0,uniqu:[4,6],unit:[2,14,19],unittest:0,univers:8,unk:[13,18],unknown:13,unless:3,unload:12,unload_model:12,unmodifi:13,unnecessari:[0,3],unpc:[4,6],unset:3,unshar:20,until:[13,18],unwieldi:3,updat:[7,9,12,13,17,20],update_finish:13,update_learning_r:17,update_n_src_word:9,update_vocab:17,upgrad:7,upon:1,upper:3,url:[7,8,21],url_root:16,usag:[5,6,15,16,17,18],use:[0,3,4,5,6,7,9,11,12,13,14,15,17,18,19,20],used:[2,3,5,9,10,11,12,13,14,15,17,18],useful:[1,9],user:[4,5,6,7,9,12],uses:[0,3,5,6,7,11,13,17,20],using:[0,1,2,3,4,5,8,11,12,13,15,17,18,19,20],usual:7,uszkoreit:21,util:[2,4,5,9,10,14],valid:[5,9,15,17,18,20],valid_batch_s:[5,17],valid_it:9,valid_loss:9,valid_loss_md:9,valid_step:[5,9,17],valu:[2,3,4,5,9,11,12,13,15,17,18],variabl:[3,4,7,10,13],variat:0,variou:[4,14],vaswani:21,vaswanispujgkp17:0,vector:[11,14,17],venv:7,verbos:[13,17,18],veri:[0,14,18],versatil:14,version:[1,5,12,13,19],via:[11,21],vinyal:21,virtual:7,visit:0,visual:[14,17],vocab:[4,5,6,9,13,14,20],vocab_path:[15,17,18],vocab_s:[13,17,19],vocab_sample_queue_s:15,vocab_size_multipl:[17,18],vocabs_dict:10,vocabulari:[3,4,5,9,14,15,17,18,19,20],vsp:[11,21],wai:[3,5,13],wait:[3,5],wang:21,want:[3,18],warmup:17,warmup_step:[5,17],warn:[10,15,17,18],websit:[5,19],weight:[2,3,11,17,18],weight_decai:17,weighted_sampl:17,well:[0,6,17],wget:[4,5,19,20],what:[3,6,9,12],when:[0,3,8,10,11,13,15,17,18,19,20],where:[2,3,6,7,11,13,14,15,17,18],wherea:[13,17],whether:[9,11,12,13,15,17,18],which:[1,3,4,11,13,14,17,19,20],whl:7,whole:13,whose:18,why:2,wiki:17,wikipedia:17,window:[15,17,18],with_align:9,within:[2,11,12],without:[0,7,17],wojciech:21,wolfgang:21,word2vec:[17,18],word:[2,11,13,14,15,17,18],word_align:13,word_lut:11,word_padding_idx:11,word_ratio_threshold:[15,17,18],word_vec_s:11,word_vocab_s:11,work:[0,3,7,13,17,20],workflow:8,workstat:[],world_siz:[5,17,20],would:[3,11,13,17,19],wouldn:4,wpdn18:[15,17,18,21],wrap:[11,12],wrapper:[4,5,9],writabl:3,write:[3,5,9,19],writer:9,wsc:[13,21],www:17,xavier_uniform:17,xent:9,xinyi:21,xiong:21,xvzf:19,xx_side_in:19,xx_side_out:19,xx_stream:19,xzvf:19,yaml:[4,5,15,17,18,20],year:[4,8,19],yet:13,yield:10,yml:[0,5,20],yonghui:21,yoon:8,you:[0,3,4,5,7,11,17,18,19,20,21],your:[0,3,4,6,7,19,20],your_config:[],your_path:5,your_project_name_is_your_account:7,your_venv_nam:7,your_vevn_nam:7,yourself:[8,19],yuan:21,yuntian:8,yyi:3,zaremba:21,zero:[3,9,11,13,15,17,18],zero_grad:9,zhang:21,zhifeng:21,ziemski:19,zihang:21,zip:19,zquez:8,zxs18:[11,21]},titles:["Contributors","About MAMMOTH","Attention Bridge","Config-config Tool","MAMMOTH Sharing Schemes","Training MAMMOTH 101","Contents","Installation","Overview","Framework","Data Loaders","Modules","Server","Translation","Component-level Modularity","Build Vocab","Server","Train","Translate","Prepare Data","Quickstart","References"],titleterms:{"case":1,"class":13,"function":1,"while":1,Are:1,The:3,about:1,actual:3,adapt:[3,17],adapter_config:3,ae_path:3,ae_transform:3,algorithm:1,align:17,allocate_devic:3,altern:3,anatomi:14,ani:1,approach:1,argument:16,attent:[2,11,17],autoencod:3,beam:18,between:1,bridg:[2,11,14,17],build:15,can:1,capabl:1,citat:8,cluster_languag:3,command:[3,4],common:[15,17,18],complete_language_pair:3,compon:14,config:3,config_al:3,config_config:3,configur:[5,15,17,18,20],content:6,contribut:1,contributor:0,core:12,corpora:3,corpora_schedul:3,creation:5,custom:14,data:[5,10,15,17,18,19,20],dataset:[4,10],dec_sharing_group:3,decod:[4,11,13,17,18],denois:[15,17,18],develop:1,direct:19,directori:5,distanc:3,distance_matrix:3,docstr:0,document:1,doe:1,domain:1,download:[5,19],dynam:17,effici:18,embed:[11,17,18],enc_sharing_group:3,encod:[4,11,17],enhanc:1,environ:5,europarl:19,featur:[1,17],feedforwardattentionbridgelay:2,filter:[15,17,18],flexibl:14,framework:9,fulli:4,further:20,futur:1,gener:17,get:19,group:3,guidelin:0,has:1,how:1,infer:4,inferfeat:[15,17,18],initi:17,innov:1,input:3,instal:[7,8,20],introduc:1,issu:1,its:1,job:5,kei:[1,3],languag:3,level:[3,14],leverag:1,linattentionbridgelay:2,line:3,load:10,loader:10,log:[17,18],loss:9,lumi:7,made:1,mahti:7,mammoth:[1,4,5,20],manual:3,matrix:3,model:[5,9,12,14,17,18,19],modif:1,modul:11,modular:[4,14],n_gpus_per_nod:3,n_group:3,n_node:3,name:16,note:4,onmttok:[15,17,18],opennmt:1,optim:[9,17],opu:19,origin:1,other:3,overrid:3,overview:[4,8],own:1,parallel:14,paramet:[3,14],pars:19,particularli:1,path:19,penalti:18,perceiverattentionbridgelay:2,plan:1,prepar:[5,19,20],project:1,prune:[17,18],puhti:7,quickstart:[19,20],random:18,rate:17,read:20,refer:21,relationship:1,relev:19,remove_temporary_kei:3,report:1,repositori:1,reproduc:[15,17,18],run:7,sampl:18,scheme:4,score:13,search:18,sentencepiec:[5,19],separ:4,server:[12,16],set:19,set_transform:3,setup:[4,5],share:[3,4,14],sharing_group:3,shot:19,simpleattentionbridgelay:2,specif:1,specifi:3,src_path:3,stage:3,start:20,statu:1,step:[5,19,20],strategi:13,structur:14,sublay:11,submiss:5,subword:[15,17,18],suit:1,supervis:19,switchout:[15,17,18],system:4,task:[5,15,17,18],test:19,tgt_path:3,than:3,token:19,token_drop:[15,17,18],token_mask:[15,17,18],tool:3,top:3,train:[4,5,17,19,20],trainer:9,transform:[3,15,17,18],transformerattentionbridgelay:2,translat:[13,18,19,20],translation_config:3,translation_config_dir:3,trick:18,type:17,uniqu:1,unpc:19,unshar:4,usag:3,use:1,use_introduce_at_training_step:3,use_src_lang_token:3,use_weight:3,user:1,uses:1,valid:19,variabl:[5,19],vocab:[10,15,17,18,19],well:1,what:1,where:1,yaml:3,your:5,zero:19,zero_shot:3}}) \ No newline at end of file