gen_anchors_bert.py

# coding=utf-8
# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


"""This script was copied and modified from https://github.com/huggingface/pytorch-pretrained-BERT"""
"""Note: this version of the code doesn't handle words that are split into different subwords"""

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from tqdm import tqdm
import argparse
import numpy as np
import collections
import logging
import json
import re
import os

import torch
from torch.utils.data import TensorDataset, DataLoader, SequentialSampler
from torch.utils.data.distributed import DistributedSampler

from pytorch_pretrained_bert.tokenization import BertTokenizer
from pytorch_pretrained_bert.modeling import BertModel

from allennlp.data import vocabulary

logging.basicConfig(format = '%(asctime)s - %(levelname)s - %(name)s -   %(message)s', 
                    datefmt = '%m/%d/%Y %H:%M:%S',
                    level = logging.INFO)
logger = logging.getLogger(__name__)


class InputExample(object):

    def __init__(self, unique_id, text_a, text_b):
        self.unique_id = unique_id
        self.text_a = text_a
        self.text_b = text_b


class InputFeatures(object):
    """A single set of features of data."""

    def __init__(self, unique_id, tokens, input_ids, input_mask, input_type_ids):
        self.unique_id = unique_id
        self.tokens = tokens
        self.input_ids = input_ids
        self.input_mask = input_mask
        self.input_type_ids = input_type_ids


def convert_examples_to_features(examples, seq_length, tokenizer):
    """Loads a data file into a list of `InputFeature`s."""

    features = []
    for (ex_index, example) in enumerate(examples):
        tokens_a = tokenizer.tokenize(example.text_a)

        tokens_b = None
        if example.text_b:
            tokens_b = tokenizer.tokenize(example.text_b)

        if tokens_b:
            # Modifies `tokens_a` and `tokens_b` in place so that the total
            # length is less than the specified length.
            # Account for [CLS], [SEP], [SEP] with "- 3"
            _truncate_seq_pair(tokens_a, tokens_b, seq_length - 3)
        else:
            # Account for [CLS] and [SEP] with "- 2"
            if len(tokens_a) > seq_length - 2:
                tokens_a = tokens_a[0:(seq_length - 2)]

        # The convention in BERT is:
        # (a) For sequence pairs:
        #  tokens:   [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP]
        #  type_ids:   0   0  0    0    0     0      0   0    1  1  1   1  1   1
        # (b) For single sequences:
        #  tokens:   [CLS] the dog is hairy . [SEP]
        #  type_ids:   0   0   0   0  0     0   0
        #
        # Where "type_ids" are used to indicate whether this is the first
        # sequence or the second sequence. The embedding vectors for `type=0` and
        # `type=1` were learned during pre-training and are added to the wordpiece
        # embedding vector (and position vector). This is not *strictly* necessary
        # since the [SEP] token unambigiously separates the sequences, but it makes
        # it easier for the model to learn the concept of sequences.
        #
        # For classification tasks, the first vector (corresponding to [CLS]) is
        # used as as the "sentence vector". Note that this only makes sense because
        # the entire model is fine-tuned.
        tokens = []
        input_type_ids = []
        tokens.append("[CLS]")
        input_type_ids.append(0)
        for token in tokens_a:
            tokens.append(token)
            input_type_ids.append(0)
        tokens.append("[SEP]")
        input_type_ids.append(0)

        if tokens_b:
            for token in tokens_b:
                tokens.append(token)
                input_type_ids.append(1)
            tokens.append("[SEP]")
            input_type_ids.append(1)

        input_ids = tokenizer.convert_tokens_to_ids(tokens)

        # The mask has 1 for real tokens and 0 for padding tokens. Only real
        # tokens are attended to.
        input_mask = [1] * len(input_ids)

        # Zero-pad up to the sequence length.
        while len(input_ids) < seq_length:
            input_ids.append(0)
            input_mask.append(0)
            input_type_ids.append(0)

        assert len(input_ids) == seq_length
        assert len(input_mask) == seq_length
        assert len(input_type_ids) == seq_length

        if ex_index < 5:
            logger.info("*** Example ***")
            logger.info("unique_id: %s" % (example.unique_id))
            logger.info("tokens: %s" % " ".join([str(x) for x in tokens]))
            logger.info("input_ids: %s" % " ".join([str(x) for x in input_ids]))
            logger.info("input_mask: %s" % " ".join([str(x) for x in input_mask]))
            logger.info(
                "input_type_ids: %s" % " ".join([str(x) for x in input_type_ids]))

        features.append(
            InputFeatures(
                unique_id=example.unique_id,
                tokens=tokens,
                input_ids=input_ids,
                input_mask=input_mask,
                input_type_ids=input_type_ids))
    return features


def _truncate_seq_pair(tokens_a, tokens_b, max_length):
    """Truncates a sequence pair in place to the maximum length."""

    # This is a simple heuristic which will always truncate the longer sequence
    # one token at a time. This makes more sense than truncating an equal percent
    # of tokens from each, since if one sequence is very short then each token
    # that's truncated likely contains more information than a longer sequence.
    while True:
        total_length = len(tokens_a) + len(tokens_b)
        if total_length <= max_length:
            break
        if len(tokens_a) > len(tokens_b):
            tokens_a.pop()
        else:
            tokens_b.pop()


def read_examples(input_file):
    """Read a list of `InputExample`s from an input file."""
    examples = []
    unique_id = 0
    with open(input_file, "r", encoding='utf-8') as reader:
        while True:
            line = reader.readline()
            if not line:
                break
            line = line.strip()
            text_a = None
            text_b = None
            m = re.match(r"^(.*) \|\|\| (.*)$", line)
            if m is None:
                text_a = line
            else:
                text_a = m.group(1)
                text_b = m.group(2)
            examples.append(
                InputExample(unique_id=unique_id, text_a=text_a, text_b=text_b))
            unique_id += 1
    return examples

def save_embeds(file_path, embeds, vocab, num_occurrences, emb_dim):
    # Don't include words not in the text.
    n_tokens = len(np.nonzero(num_occurrences)[0])
    with open(file_path, 'w') as f:
        f.write('%d %d\n' % (n_tokens, emb_dim))
        for i in range(embeds.shape[0]):
            if num_occurrences[i] == 0:
                continue

            token = vocab.get_token_from_index(i)
            to_dump = token + ' ' + ' '.join([str(v) for v in embeds[i, :]]) + '\n'
            f.write(to_dump)


def main():
    parser = argparse.ArgumentParser()

    ## Required parameters
    parser.add_argument("--input_file", default=None, type=str, required=True)
    parser.add_argument("--out_dir", default=None, type=str, required=True)
    parser.add_argument("--bert_model", default=None, type=str, required=True,
                        help="Bert pre-trained model selected in the list: bert-base-uncased, "
                             "bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese.")

    ## Other parameters
    parser.add_argument("--do_lower_case", action='store_true', help="Set this flag if you are using an uncased model.")
    parser.add_argument("--layers", default="-1,-2,-3,-4", type=str)
    parser.add_argument("--max_seq_length", default=128, type=int,
                        help="The maximum total input sequence length after WordPiece tokenization. Sequences longer "
                            "than this will be truncated, and sequences shorter than this will be padded.")
    parser.add_argument("--batch_size", default=32, type=int, help="Batch size for predictions.")
    parser.add_argument("--local_rank",
                        type=int,
                        default=-1,
                        help = "local_rank for distributed training on gpus")
    parser.add_argument("--no_cuda",
                        action='store_true',
                        help="Whether not to use CUDA when available")
    parser.add_argument(
        '-d', '--emb_dim', type=int, default=768, help="Embeddings size")
    parser.add_argument(
        '--vocab_file',
        type=str,
        default='vocabs/en_50k.vocab',
        help=
        "Path to vocab file with tokens (one per line) to include in output. Should also include <UNK> token. Can use $l as a placeholder for language"
    )

    args = parser.parse_args()

    vocab = vocabulary.Vocabulary()
    vocab.set_from_file(args.vocab_file, oov_token='<UNK>')
    print("Loaded vocabulary of size {}".format(vocab.get_vocab_size()))

    if args.local_rank == -1 or args.no_cuda:
        device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
        n_gpu = torch.cuda.device_count()
    else:
        device = torch.device("cuda", args.local_rank)
        n_gpu = 1
        # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
        torch.distributed.init_process_group(backend='nccl')
    logger.info("device: {} n_gpu: {} distributed training: {}".format(device, n_gpu, bool(args.local_rank != -1)))

    layer_indexes = [int(x) for x in args.layers.split(",")]

    tokenizer = BertTokenizer.from_pretrained(args.bert_model, do_lower_case=args.do_lower_case)

    examples = read_examples(args.input_file)

    features = convert_examples_to_features(
        examples=examples, seq_length=args.max_seq_length, tokenizer=tokenizer)

    unique_id_to_feature = {}
    for feature in features:
        unique_id_to_feature[feature.unique_id] = feature

    model = BertModel.from_pretrained(args.bert_model)
    model.to(device)

    if args.local_rank != -1:
        model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.local_rank],
                                                          output_device=args.local_rank)
    elif n_gpu > 1:
        model = torch.nn.DataParallel(model)

    all_input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)
    all_input_mask = torch.tensor([f.input_mask for f in features], dtype=torch.long)
    all_example_index = torch.arange(all_input_ids.size(0), dtype=torch.long)

    eval_data = TensorDataset(all_input_ids, all_input_mask, all_example_index)
    if args.local_rank == -1:
        eval_sampler = SequentialSampler(eval_data)
    else:
        eval_sampler = DistributedSampler(eval_data)
    eval_dataloader = DataLoader(eval_data, sampler=eval_sampler, batch_size=args.batch_size)

    num_occurrences = [0] * vocab.get_vocab_size()
    anchors = {}
    norms = {}
    total_words = 0
    for l in layer_indexes:
        norms[l] = 0.0
        anchors[l] = np.zeros(
            shape=(vocab.get_vocab_size(), args.emb_dim))

    oov_ind = vocab.get_token_index(vocab._oov_token)
    model.eval()
    for input_ids, input_mask, example_indices in tqdm(eval_dataloader):
            input_ids = input_ids.to(device)
            input_mask = input_mask.to(device)

            all_encoder_layers, _ = model(input_ids, token_type_ids=None, attention_mask=input_mask)
            all_encoder_layers = all_encoder_layers

            for b, example_index in enumerate(example_indices):
                feature = features[example_index.item()]

                for (i, token) in enumerate(feature.tokens):
                    all_layers = []
                    w_id = vocab.get_token_index(token)
                    if w_id == oov_ind:
                        continue

                    n = num_occurrences[w_id]
                    for (j, layer_index) in enumerate(layer_indexes):
                        layer_output = all_encoder_layers[int(layer_index)].detach().cpu().numpy()
                        layer_output = layer_output[b]
                        layers = collections.OrderedDict()
                        l = layer_index
                        values = layer_output[i]

                        anchors[l][
                            w_id, :] = anchors[l][w_id, :] * (
                                n / (n + 1)) + values[:] / (n + 1)
                        norm = np.linalg.norm(values[:])
                        norms[l] = norms[l] * (total_words / (total_words +
                                               1)) + norm / (total_words + 1)

                    total_words += 1
                    num_occurrences[w_id] += 1

    os.makedirs(args.out_dir, exist_ok=True)
    norm_dict = {}
    print('Saving outputs to {}'.format(args.out_dir))
    for l in tqdm(layer_indexes):
        norm_key = 'avg_norm_layer_{}'.format(l)
        norm_dict[norm_key] = norms[l]
        file_path = os.path.join(args.out_dir, 'avg_embeds_{}.txt'.format(l))
        save_embeds(file_path, anchors[l], vocab, num_occurrences, args.emb_dim)

    norm_dict['occurrences'] = num_occurrences
    file_path = os.path.join(args.out_dir, 'norms.json')
    json.dump(norm_dict, open(file_path, 'w'))
    
if __name__ == "__main__":
    main()