490f949617 | ||
---|---|---|
.. | ||
config | ||
scripts | ||
src | ||
README.md | ||
__init__.py | ||
apply_bpe_encoding.py | ||
cornell_dialog.py | ||
eval.py | ||
gigaword.py | ||
news_crawl.py | ||
requirements.txt | ||
tokenize_corpus.py | ||
train.py | ||
weights_average.py |
README.md
- MASS: Masked Sequence to Sequence Pre-training for Language Generation Description
- Model architecture
- Dataset
- Features
- Script description
- Model description
- Environment Requirements
- Get started
- Description of random situation
- others
- ModelZoo Homepage
MASS: Masked Sequence to Sequence Pre-training for Language Generation Description
MASS: Masked Sequence to Sequence Pre-training for Language Generation was released by MicroSoft in June 2019.
BERT(Devlin et al., 2018) have achieved SOTA in natural language understanding area by pre-training the encoder part of Transformer(Vaswani et al., 2017) with masked rich-resource text. Likewise, GPT(Raddford et al., 2018) pre-trains the decoder part of Transformer with masked(encoder inputs are masked) rich-resource text. Both of them build a robust language model by pre-training with masked rich-resource text.
Inspired by BERT, GPT and other language models, MicroSoft addressed MASS: Masked Sequence to Sequence Pre-training for Language Generation which combines BERT's and GPT's idea. MASS has an important parameter k, which controls the masked fragment length. BERT and GPT are specicl case when k equals to 1 and sentence length.
Paper: Song, Kaitao, Xu Tan, Tao Qin, Jianfeng Lu and Tie-Yan Liu. “MASS: Masked Sequence to Sequence Pre-training for Language Generation.” ICML (2019).
Model architecture
The overall network architecture of MASS is shown below, which is Transformer(Vaswani et al., 2017):
MASS is consisted of 6-layer encoder and 6-layer decoder with 1024 embedding/hidden size, and 4096 intermediate size between feed forward network which has two full connection layers.
Dataset
Dataset used:
- monolingual English data from News Crawl dataset(WMT 2019) for pre-training.
- Gigaword Corpus(Graff et al., 2003) for Text Summarization.
- Cornell movie dialog corpus(DanescuNiculescu-Mizil & Lee, 2011).
Details about those dataset could be found in MASS: Masked Sequence to Sequence Pre-training for Language Generation.
Features
Mass is designed to jointly pre train encoder and decoder to complete the task of language generation. First of all, through a sequence to sequence framework, mass only predicts the blocked token, which forces the encoder to understand the meaning of the unshielded token, and encourages the decoder to extract useful information from the encoder. Secondly, by predicting the continuous token of the decoder, the decoder can build better language modeling ability than only predicting discrete token. Third, by further shielding the input token of the decoder which is not shielded in the encoder, the decoder is encouraged to extract more useful information from the encoder side, rather than using the rich information in the previous token.
Script description
MASS script and code structure are as follow:
├── mass
├── README.md // Introduction of MASS model.
├── config
│ ├──config.py // Configuration instance definition.
│ ├──config.json // Configuration file.
├── src
│ ├──dataset
│ ├──bi_data_loader.py // Dataset loader for fine-tune or inferring.
│ ├──mono_data_loader.py // Dataset loader for pre-training.
│ ├──language_model
│ ├──noise_channel_language_model.p // Noisy channel language model for dataset generation.
│ ├──mass_language_model.py // MASS language model according to MASS paper.
│ ├──loose_masked_language_model.py // MASS language model according to MASS released code.
│ ├──masked_language_model.py // Masked language model according to MASS paper.
│ ├──transformer
│ ├──create_attn_mask.py // Generate mask matrix to remove padding positions.
│ ├──transformer.py // Transformer model architecture.
│ ├──encoder.py // Transformer encoder component.
│ ├──decoder.py // Transformer decoder component.
│ ├──self_attention.py // Self-Attention block component.
│ ├──multi_head_attention.py // Multi-Head Self-Attention component.
│ ├──embedding.py // Embedding component.
│ ├──positional_embedding.py // Positional embedding component.
│ ├──feed_forward_network.py // Feed forward network.
│ ├──residual_conn.py // Residual block.
│ ├──beam_search.py // Beam search decoder for inferring.
│ ├──transformer_for_infer.py // Use Transformer to infer.
│ ├──transformer_for_train.py // Use Transformer to train.
│ ├──utils
│ ├──byte_pair_encoding.py // Apply BPE with subword-nmt.
│ ├──dictionary.py // Dictionary.
│ ├──loss_moniter.py // Callback of monitering loss during training step.
│ ├──lr_scheduler.py // Learning rate scheduler.
│ ├──ppl_score.py // Perplexity score based on N-gram.
│ ├──rouge_score.py // Calculate ROUGE score.
│ ├──load_weights.py // Load weights from a checkpoint or NPZ file.
│ ├──initializer.py // Parameters initializer.
├── vocab
│ ├──all.bpe.codes // BPE codes table(this file should be generated by user).
│ ├──all_en.dict.bin // Learned vocabulary file(this file should be generated by user).
├── scripts
│ ├──run.sh // Train & evaluate model script.
│ ├──learn_subword.sh // Learn BPE codes.
│ ├──stop_training.sh // Stop training.
├── requirements.txt // Requirements of third party package.
├── train.py // Train API entry.
├── eval.py // Infer API entry.
├── tokenize_corpus.py // Corpus tokenization.
├── apply_bpe_encoding.py // Applying bpe encoding.
├── weights_average.py // Average multi model checkpoints to NPZ format.
├── news_crawl.py // Create News Crawl dataset for pre-training.
├── gigaword.py // Create Gigaword Corpus.
├── cornell_dialog.py // Create Cornell Movie Dialog dataset for conversation response.
Data Preparation
The data preparation of a natural language processing task contains data cleaning, tokenization, encoding and vocabulary generation steps.
In our experiments, using Byte Pair Encoding(BPE) could reduce size of vocabulary, and relieve the OOV influence effectively.
Vocabulary could be created using src/utils/dictionary.py
with text dictionary which is learnt from BPE.
For more detail about BPE, please refer to Subword-nmt lib or paper.
In our experiments, vocabulary was learned based on 1.9M sentences from News Crawl Dataset, size of vocabulary is 45755.
Here, we have a brief introduction of data preparation scripts.
Tokenization
Using tokenize_corpus.py
could tokenize corpus whose text files are in format of .txt
.
Major parameters in tokenize_corpus.py
:
--corpus_folder: Corpus folder path, if multi-folders are provided, use ',' split folders.
--output_folder: Output folder path.
--tokenizer: Tokenizer to be used, nltk or jieba, if nltk is not installed fully, use jieba instead.
--pool_size: Processes pool size.
Sample code:
python tokenize_corpus.py --corpus_folder /{path}/corpus --output_folder /{path}/tokenized_corpus --tokenizer {nltk|jieba} --pool_size 16
Byte Pair Encoding
After tokenization, BPE is applied to tokenized corpus with provided all.bpe.codes
.
Apply BPE script can be found in apply_bpe_encoding.py
.
Major parameters in apply_bpe_encoding.py
:
--codes: BPE codes file.
--src_folder: Corpus folders.
--output_folder: Output files folder.
--prefix: Prefix of text file in `src_folder`.
--vocab_path: Generated vocabulary output path.
--threshold: Filter out words that frequency is lower than threshold.
--processes: Size of process pool (to accelerate). Default: 2.
Sample code:
python tokenize_corpus.py --codes /{path}/all.bpe.codes \
--src_folder /{path}/tokenized_corpus \
--output_folder /{path}/tokenized_corpus/bpe \
--prefix tokenized \
--vocab_path /{path}/vocab_en.dict.bin
--processes 32
Build Vocabulary
Support that you want to create a new vocabulary, there are two options:
- Learn BPE codes from scratch, and create vocabulary with multi vocabulary files from
subword-nmt
. - Create from an existing vocabulary file which lines in the format of
word frequency
. - Optional, Create a small vocabulary based on
vocab/all_en.dict.bin
with method ofshink
fromsrc/utils/dictionary.py
. - Persistent vocabulary to
vocab
folder with methodpersistence()
.
Major interface of src/utils/dictionary.py
are as follow:
shrink(self, threshold=50)
: Shrink the size of vocabulary by filter out words frequency is lower than threshold. It returns a new vocabulary.load_from_text(cls, filepaths: List[str])
: Load existed text vocabulary which lines in the format ofword frequency
.load_from_persisted_dict(cls, filepath)
: Load from a persisted binary vocabulary which was saved by callingpersistence()
method.persistence(self, path)
: Save vocabulary object to binary file.
Sample code:
from src.utils import Dictionary
vocabulary = Dictionary.load_from_persisted_dict("vocab/all_en.dict.bin")
tokens = [1, 2, 3, 4, 5]
# Convert ids to symbols.
print([vocabulary[t] for t in tokens])
sentence = ["Hello", "world"]
# Convert symbols to ids.
print([vocabulary.index[s] for s in sentence])
For more detail, please refer to the source file.
Generate Dataset
As mentioned above, three corpus are used in MASS mode, dataset generation scripts for them are provided.
News Crawl Corpus
Script can be found in news_crawl.py
.
Major parameters in news_crawl.py
:
Note that please provide `--existed_vocab` or `--dict_folder` at least one.
A new vocabulary would be created in `output_folder` when pass `--dict_folder`.
--src_folder: Corpus folders.
--existed_vocab: Optional, persisted vocabulary file.
--mask_ratio: Ratio of mask.
--output_folder: Output dataset files folder path.
--max_len: Maximum sentence length. If a sentence longer than `max_len`, then drop it.
--suffix: Optional, suffix of generated dataset files.
--processes: Optional, size of process pool (to accelerate). Default: 2.
Sample code:
python news_crawl.py --src_folder /{path}/news_crawl \
--existed_vocab /{path}/mass/vocab/all_en.dict.bin \
--mask_ratio 0.5 \
--output_folder /{path}/news_crawl_dataset \
--max_len 32 \
--processes 32
Gigaword Corpus
Script can be found in gigaword.py
.
Major parameters in gigaword.py
:
--train_src: Train source file path.
--train_ref: Train reference file path.
--test_src: Test source file path.
--test_ref: Test reference file path.
--existed_vocab: Persisted vocabulary file.
--output_folder: Output dataset files folder path.
--noise_prob: Optional, add noise prob. Default: 0.
--max_len: Optional, maximum sentence length. If a sentence longer than `max_len`, then drop it. Default: 64.
--format: Optional, dataset format, "mindrecord" or "tfrecord". Default: "tfrecord".
Sample code:
python gigaword.py --train_src /{path}/gigaword/train_src.txt \
--train_ref /{path}/gigaword/train_ref.txt \
--test_src /{path}/gigaword/test_src.txt \
--test_ref /{path}/gigaword/test_ref.txt \
--existed_vocab /{path}/mass/vocab/all_en.dict.bin \
--noise_prob 0.1 \
--output_folder /{path}/gigaword_dataset \
--max_len 64
Cornell Movie Dialog Corpus
Script can be found in cornell_dialog.py
.
Major parameters in cornell_dialog.py
:
--src_folder: Corpus folders.
--existed_vocab: Persisted vocabulary file.
--train_prefix: Train source and target file prefix. Default: train.
--test_prefix: Test source and target file prefix. Default: test.
--output_folder: Output dataset files folder path.
--max_len: Maximum sentence length. If a sentence longer than `max_len`, then drop it.
--valid_prefix: Optional, Valid source and target file prefix. Default: valid.
Sample code:
python cornell_dialog.py --src_folder /{path}/cornell_dialog \
--existed_vocab /{path}/mass/vocab/all_en.dict.bin \
--train_prefix train \
--test_prefix test \
--noise_prob 0.1 \
--output_folder /{path}/cornell_dialog_dataset \
--max_len 64
Configuration
Json file under the path config/
is the template configuration file.
Almost all of the options and arguments needed could be assigned conveniently, including the training platform, configurations of dataset and model, arguments of optimizer etc. Optional features such as loss scale and checkpoint are also available by setting the options correspondingly.
For more detailed information about the attributes, refer to the file config/config.py
.
Training & Evaluation process
For training a model, the shell script run.sh
is all you need. In this scripts, the environment variable is set and the training script train.py
under mass
is executed.
You may start a task training with single device or multiple devices by assigning the options and run the command in bash:
sh run.sh [--options]
The usage is shown as bellow:
Usage: run.sh [-h, --help] [-t, --task <CHAR>] [-n, --device_num <N>]
[-i, --device_id <N>] [-j, --hccl_json <FILE>]
[-c, --config <FILE>] [-o, --output <FILE>]
[-v, --vocab <FILE>]
options:
-h, --help show usage
-t, --task select task: CHAR, 't' for train and 'i' for inference".
-n, --device_num device number used for training: N, default is 1.
-i, --device_id device id used for training with single device: N, 0<=N<=7, default is 0.
-j, --hccl_json rank table file used for training with multiple devices: FILE.
-c, --config configuration file as shown in the path 'mass/config': FILE.
-o, --output assign output file of inference: FILE.
-v, --vocab set the vocabulary"
Notes: Be sure to assign the hccl_json file while running a distributed-training.
The command followed shows a example for training with 2 devices.
sh run.sh --task t --device_num 2 --hccl_json /{path}/rank_table.json --config /{path}/config.json
ps. Discontinuous device id is not supported in run.sh
at present, device id in rank_table.json
must start from 0.
If use a single chip, it would be like this:
sh run.sh --task t --device_num 1 --device_id 0 --config /{path}/config.json
Weights average
python weights_average.py --input_files your_checkpoint_list --output_file model.npz
The input_files is a list of you checkpoints file. To use model.npz as the weights, add its path in config.json at "existed_ckpt".
{
...
"checkpoint_options": {
"existed_ckpt": "/xxx/xxx/model.npz",
"save_ckpt_steps": 1000,
...
},
...
}
Learning rate scheduler
Two learning rate scheduler are provided in our model:
LR scheduler could be config in config/config.json
.
For Polynomial decay scheduler, config could be like:
{
...
"learn_rate_config": {
"optimizer": "adam",
"lr": 1e-4,
"lr_scheduler": "poly",
"poly_lr_scheduler_power": 0.5,
"decay_steps": 10000,
"warmup_steps": 2000,
"min_lr": 1e-6
},
...
}
For Inverse square root scheduler, config could be like:
{
...
"learn_rate_config": {
"optimizer": "adam",
"lr": 1e-4,
"lr_scheduler": "isr",
"decay_start_step": 12000,
"warmup_steps": 2000,
"min_lr": 1e-6
},
...
}
More detail about LR scheduler could be found in src/utils/lr_scheduler.py
.
Model description
The MASS network is implemented by Transformer, which has multi-encoder layers and multi-decoder layers. For pre-training, we use the Adam optimizer and loss-scale to get the pre-trained model. During fine-turning, we fine-tune this pre-trained model with different dataset according to different tasks. During testing, we use the fine-turned model to predict the result, and adopt a beam search algorithm to get the most possible prediction results.
Performance
Results
Fine-Tuning on Text Summarization
The comparisons between MASS and two other pre-training methods in terms of ROUGE score on the text summarization task with 3.8M training data are as follows:
Method | RG-1(F) | RG-2(F) | RG-L(F) |
---|---|---|---|
MASS | Ongoing | Ongoing | Ongoing |
Fine-Tuning on Conversational ResponseGeneration
The comparisons between MASS and other baseline methods in terms of PPL on Cornell Movie Dialog corpus are as follows:
Method | Data = 10K | Data = 110K |
---|---|---|
MASS | Ongoing | Ongoing |
Training Performance
Parameters | Masked Sequence to Sequence Pre-training for Language Generation |
---|---|
Model Version | v1 |
Resource | Ascend 910, cpu 2.60GHz, 56cores;memory, 314G |
uploaded Date | 05/24/2020 |
MindSpore Version | 0.2.0 |
Dataset | News Crawl 2007-2017 English monolingual corpus, Gigaword corpus, Cornell Movie Dialog corpus |
Training Parameters | Epoch=50, steps=XXX, batch_size=192, lr=1e-4 |
Optimizer | Adam |
Loss Function | Label smoothed cross-entropy criterion |
outputs | Sentence and probability |
Loss | Lower than 2 |
Accuracy | For conversation response, ppl=23.52, for text summarization, RG-1=29.79. |
Speed | 611.45 sentences/s |
Total time | --/-- |
Params (M) | 44.6M |
Checkpoint for Fine tuning | ---Mb, --, A link |
Model for inference | ---Mb, --, A link |
Scripts | A link |
Inference Performance
Parameters | Masked Sequence to Sequence Pre-training for Language Generation |
---|---|
Model Version | V1 |
Resource | Huawei 910 |
uploaded Date | 05/24/2020 |
MindSpore Version | 0.2.0 |
Dataset | Gigaword corpus, Cornell Movie Dialog corpus |
batch_size | --- |
outputs | Sentence and probability |
Accuracy | ppl=23.52 for conversation response, RG-1=29.79 for text summarization. |
Speed | ---- sentences/s |
Total time | --/-- |
Model for inference | ---Mb, --, A link |
Environment Requirements
Platform
- Hardware(Ascend)
- Prepare hardware environment with Ascend processor. If you want to try Ascend, please send the application form to ascend@huawei.com. Once approved, you could get the resources for trial.
- Framework
- For more information, please check the resources below:
Requirements
nltk
numpy
subword-nmt
rouge
https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/network_migration.html
Get started
MASS pre-trains a sequence to sequence model by predicting the masked fragments in an input sequence. After this, downstream tasks including text summarization and conversation response are candidated for fine-tuning the model and for inference. Here we provide a practice example to demonstrate the basic usage of MASS for pre-training, fine-tuning a model, and the inference process. The overall process is as follows:
- Download and process the dataset.
- Modify the
config.json
to config the network. - Run a task for pre-training and fine-tuning.
- Perform inference and validation.
Pre-training
For pre-training a model, config the options in config.json
firstly:
- Assign the
pre_train_dataset
underdataset_config
node to the dataset path. - Choose the optimizer('momentum/adam/lamb' is available).
- Assign the 'ckpt_prefix' and 'ckpt_path' under
checkpoint_path
to save the model files. - Set other arguments including dataset configurations and network configurations.
- If you have a trained model already, assign the
existed_ckpt
to the checkpoint file.
Run the shell script run.sh
as followed:
sh run.sh -t t -n 1 -i 1 -c /mass/config/config.json
Get the log and output files under the path ./run_mass_*/
, and the model file under the path assigned in the config/config.json
file.
Fine-tuning
For fine-tuning a model, config the options in config.json
firstly:
- Assign the
fine_tune_dataset
underdataset_config
node to the dataset path. - Assign the
existed_ckpt
undercheckpoint_path
node to the existed model file generated by pre-training. - Choose the optimizer('momentum/adam/lamb' is available).
- Assign the
ckpt_prefix
andckpt_path
undercheckpoint_path
node to save the model files. - Set other arguments including dataset configurations and network configurations.
Run the shell script run.sh
as followed:
sh run.sh -t t -n 1 -i 1 -c config/config.json
Get the log and output files under the path ./run_mass_*/
, and the model file under the path assigned in the config/config.json
file.
Inference
If you need to use the trained model to perform inference on multiple hardware platforms, such as GPU, Ascend 910 or Ascend 310, you can refer to this Link.
For inference, config the options in config.json
firstly:
- Assign the
test_dataset
underdataset_config
node to the dataset path. - Assign the
existed_ckpt
undercheckpoint_path
node to the model file produced by fine-tuning. - Choose the optimizer('momentum/adam/lamb' is available).
- Assign the
ckpt_prefix
andckpt_path
undercheckpoint_path
node to save the model files. - Set other arguments including dataset configurations and network configurations.
Run the shell script run.sh
as followed:
sh run.sh -t i -n 1 -i 1 -c config/config.json -o {outputfile}
Description of random situation
MASS model contains dropout operations, if you want to disable dropout, please set related dropout_rate to 0 in config/config.json
.
others
The model has been validated on Ascend environment, not validated on CPU and GPU.