BERT SQuAD Example - Updates (#815)

* update to readme for pip3 upgrade, needed for tf2.4. requirement file change. initial commit for bert example update

* remove tokenization.py file

* added tokenizers and bertwordpiecetokenizer to main file

* added tokenizers to requirements file

* changes to run_onnx_squad after importing tokenizers module, to replace py file

* additional post processing tokenizer change in run_onnx_squad.py

* changes to notebook after tokenizers

* cleanup notebook output cells

* typo in readme

* formatting on py file
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

examples/python_bert_squad_example/BERT-Squad.ipynb

@@ -43,7 +43,7 @@
     "from os import path\n",
     "import sys\n",
     "\n",
-    "import tokenization\n",
+    "import tokenizers\n",
     "from run_onnx_squad import *\n",
     "\n",
     "import migraphx"
@@ -137,8 +137,7 @@
    "outputs": [],
    "source": [
     "vocab_file = os.path.join('uncased_L-12_H-768_A-12', 'vocab.txt')\n",
-    "tokenizer = tokenization.FullTokenizer(vocab_file=vocab_file,\n",
-    "                                       do_lower_case=True)"
+    "tokenizer = tokenizers.BertWordPieceTokenizer(vocab_file)"
    ]
   },
   {

examples/python_bert_squad_example/README.md

@@ -7,21 +7,25 @@ There are two ways to run the example:
 
 # Steps
 1) Install MIGraphX to your environment. Please follow the steps to build MIGraphX given at https://github.com/ROCmSoftwarePlatform/AMDMIGraphX
-2) Install the requirements file
+2) Upgrade your pip3 to latest version
+```
+pip3 install --upgrade pip 
+```
+3) Install the requirements file
 ```
 pip3 install -r requirements_bertsquad.txt
 ```
-3) Install `unzip` and fetch the uncased file (vocabulary):
+4) Install `unzip` and fetch the uncased file (vocabulary):
 ```
 apt-get install unzip
 wget -q https://storage.googleapis.com/bert_models/2018_10_18/uncased_L-12_H-768_A-12.zip
 unzip uncased_L-12_H-768_A-12.zip
 ```
-4) Get BERT ONNX model (bertsquad-10.onnx):
+5) Get BERT ONNX model (bertsquad-10.onnx):
 ```
 wget https://github.com/onnx/models/raw/master/text/machine_comprehension/bert-squad/model/bertsquad-10.onnx
 ```
-5) Run the inference, it will compile and run the model on three questions and small data provided in `inputs.json`:
+6) Run the inference, it will compile and run the model on three questions and small data provided in `inputs.json`:
 ```
 python3 bert-squad-migraphx.py
 ```

examples/python_bert_squad_example/bert-squad-migraphx.py

@@ -5,7 +5,7 @@ import os.path
 from os import path
 import sys
 
-import tokenization
+import tokenizers
 from run_onnx_squad import *
 
 import migraphx
@@ -30,8 +30,7 @@ n_best_size = 20
 max_answer_length = 30
 
 vocab_file = os.path.join('uncased_L-12_H-768_A-12', 'vocab.txt')
-tokenizer = tokenization.FullTokenizer(vocab_file=vocab_file,
-                                       do_lower_case=True)
+tokenizer = tokenizers.BertWordPieceTokenizer(vocab_file)
 
 # Use convert_examples_to_features method from run_onnx_squad to get parameters from the input
 input_ids, input_mask, segment_ids, extra_data = convert_examples_to_features(

examples/python_bert_squad_example/requirements_bertsquad.txt

-tensorflow==1.14
-onnxruntime
+tensorflow==2.4.0
+onnxruntime
\ No newline at end of file
+tokenizers
\ No newline at end of file

examples/python_bert_squad_example/run_onnx_squad.py

@@ -38,7 +38,8 @@ from timeit import default_timer as timer
 import numpy as np
 import onnxruntime as onnxrt
 import six
-import tokenization
+from tokenizers import BertWordPieceTokenizer
+from tokenizers import pre_tokenizers
 
 RawResult = collections.namedtuple("RawResult",
                                    ["unique_id", "start_logits", "end_logits"])
@@ -70,9 +71,8 @@ class SquadExample(object):
 
     def __repr__(self):
         s = []
-        s.append("qas_id: %s" % (tokenization.printable_text(self.qas_id)))
-        s.append("question_text: %s" %
-                 (tokenization.printable_text(self.question_text)))
+        s.append("qas_id: %s" % (self.qas_id))
+        s.append("question_text: %s" % (self.question_text))
         s.append("doc_tokens: [%s]" % (" ".join(self.doc_tokens)))
         if self.start_position:
             s.append("start_position: %d" % (self.start_position))
@@ -130,7 +130,7 @@ def convert_examples_to_features(examples, tokenizer, max_seq_length,
     unique_id = 0
 
     for (example_index, example) in enumerate(examples):
-        query_tokens = tokenizer.tokenize(example.question_text)
+        query_tokens = tokenizer.encode(example.question_text)
 
         if len(query_tokens) > max_query_length:
             query_tokens = query_tokens[0:max_query_length]
@@ -140,8 +140,8 @@ def convert_examples_to_features(examples, tokenizer, max_seq_length,
         all_doc_tokens = []
         for (i, token) in enumerate(example.doc_tokens):
             orig_to_tok_index.append(len(all_doc_tokens))
-            sub_tokens = tokenizer.tokenize(token)
-            for sub_token in sub_tokens:
+            sub_tokens = tokenizer.encode(token, add_special_tokens=False)
+            for sub_token in sub_tokens.tokens:
                 tok_to_orig_index.append(i)
                 all_doc_tokens.append(sub_token)
 
@@ -172,7 +172,7 @@ def convert_examples_to_features(examples, tokenizer, max_seq_length,
             segment_ids = []
             tokens.append("[CLS]")
             segment_ids.append(0)
-            for token in query_tokens:
+            for token in query_tokens.tokens:
                 tokens.append(token)
                 segment_ids.append(0)
             tokens.append("[SEP]")
@@ -192,7 +192,9 @@ def convert_examples_to_features(examples, tokenizer, max_seq_length,
             tokens.append("[SEP]")
             segment_ids.append(1)
 
-            input_ids = tokenizer.convert_tokens_to_ids(tokens)
+            input_ids = []
+            for token in tokens:
+                input_ids.append(tokenizer.token_to_id(token))
 
             # The mask has 1 for real tokens and 0 for padding tokens. Only real
             # tokens are attended to.
@@ -437,9 +439,15 @@ def get_final_text(pred_text, orig_text, do_lower_case):
     # and `pred_text`, and check if they are the same length. If they are
     # NOT the same length, the heuristic has failed. If they are the same
     # length, we assume the characters are one-to-one aligned.
-    tokenizer = tokenization.BasicTokenizer(do_lower_case=do_lower_case)
+    tokenizer = pre_tokenizers.Sequence(
+        [pre_tokenizers.Whitespace(),
+         pre_tokenizers.Punctuation()])
 
-    tok_text = " ".join(tokenizer.tokenize(orig_text))
+    tok_text = []
+    for item in tokenizer.pre_tokenize_str(orig_text):
+        tok_text.append(item[0])
+
+    tok_text = " ".join(tok_text)
 
     start_position = tok_text.find(pred_text)
     if start_position == -1:
@@ -559,8 +567,7 @@ def main():
         sess_options = onnxrt.SessionOptions()
         sess_options.session_log_verbosity_level = args.log
 
-    tokenizer = tokenization.FullTokenizer(vocab_file=args.vocab_file,
-                                           do_lower_case=True)
+    tokenizer = BertWordPieceTokenizer(vocab_file)
 
     eval_examples = read_squad_examples(input_file=args.predict_file)
     input_ids, input_mask, segment_ids, extra_data = \

examples/python_bert_squad_example/tokenization.py

-# coding=utf-8
-# Copyright 2018 The Google AI Language Team Authors.
-#
-# 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.
-"""Tokenization classes."""
-
-from __future__ import absolute_import
-from __future__ import division
-from __future__ import print_function
-
-import collections
-import re
-import unicodedata
-import six
-import tensorflow as tf
-
-
-def validate_case_matches_checkpoint(do_lower_case, init_checkpoint):
-    """Checks whether the casing config is consistent with the checkpoint name."""
-
-    # The casing has to be passed in by the user and there is no explicit check
-    # as to whether it matches the checkpoint. The casing information probably
-    # should have been stored in the bert_config.json file, but it's not, so
-    # we have to heuristically detect it to validate.
-
-    if not init_checkpoint:
-        return
-
-    m = re.match("^.*?([A-Za-z0-9_-]+)/bert_model.ckpt", init_checkpoint)
-    if m is None:
-        return
-
-    model_name = m.group(1)
-
-    lower_models = [
-        "uncased_L-24_H-1024_A-16", "uncased_L-12_H-768_A-12",
-        "multilingual_L-12_H-768_A-12", "chinese_L-12_H-768_A-12"
-    ]
-
-    cased_models = [
-        "cased_L-12_H-768_A-12", "cased_L-24_H-1024_A-16",
-        "multi_cased_L-12_H-768_A-12"
-    ]
-
-    is_bad_config = False
-    if model_name in lower_models and not do_lower_case:
-        is_bad_config = True
-        actual_flag = "False"
-        case_name = "lowercased"
-        opposite_flag = "True"
-
-    if model_name in cased_models and do_lower_case:
-        is_bad_config = True
-        actual_flag = "True"
-        case_name = "cased"
-        opposite_flag = "False"
-
-    if is_bad_config:
-        raise ValueError(
-            "You passed in `--do_lower_case=%s` with `--init_checkpoint=%s`. "
-            "However, `%s` seems to be a %s model, so you "
-            "should pass in `--do_lower_case=%s` so that the fine-tuning matches "
-            "how the model was pre-training. If this error is wrong, please "
-            "just comment out this check." %
-            (actual_flag, init_checkpoint, model_name, case_name,
-             opposite_flag))
-
-
-def convert_to_unicode(text):
-    """Converts `text` to Unicode (if it's not already), assuming utf-8 input."""
-    if six.PY3:
-        if isinstance(text, str):
-            return text
-        elif isinstance(text, bytes):
-            return text.decode("utf-8", "ignore")
-        else:
-            raise ValueError("Unsupported string type: %s" % (type(text)))
-    elif six.PY2:
-        if isinstance(text, str):
-            return text.decode("utf-8", "ignore")
-        elif isinstance(text, unicode):
-            return text
-        else:
-            raise ValueError("Unsupported string type: %s" % (type(text)))
-    else:
-        raise ValueError("Not running on Python2 or Python 3?")
-
-
-def printable_text(text):
-    """Returns text encoded in a way suitable for print or `tf.logging`."""
-
-    # These functions want `str` for both Python2 and Python3, but in one case
-    # it's a Unicode string and in the other it's a byte string.
-    if six.PY3:
-        if isinstance(text, str):
-            return text
-        elif isinstance(text, bytes):
-            return text.decode("utf-8", "ignore")
-        else:
-            raise ValueError("Unsupported string type: %s" % (type(text)))
-    elif six.PY2:
-        if isinstance(text, str):
-            return text
-        elif isinstance(text, unicode):
-            return text.encode("utf-8")
-        else:
-            raise ValueError("Unsupported string type: %s" % (type(text)))
-    else:
-        raise ValueError("Not running on Python2 or Python 3?")
-
-
-def load_vocab(vocab_file):
-    """Loads a vocabulary file into a dictionary."""
-    vocab = collections.OrderedDict()
-    index = 0
-    with tf.gfile.GFile(vocab_file, "r") as reader:
-        while True:
-            token = convert_to_unicode(reader.readline())
-            if not token:
-                break
-            token = token.strip()
-            vocab[token] = index
-            index += 1
-    return vocab
-
-
-def convert_by_vocab(vocab, items):
-    """Converts a sequence of [tokens|ids] using the vocab."""
-    output = []
-    for item in items:
-        output.append(vocab[item])
-    return output
-
-
-def convert_tokens_to_ids(vocab, tokens):
-    return convert_by_vocab(vocab, tokens)
-
-
-def convert_ids_to_tokens(inv_vocab, ids):
-    return convert_by_vocab(inv_vocab, ids)
-
-
-def whitespace_tokenize(text):
-    """Runs basic whitespace cleaning and splitting on a piece of text."""
-    text = text.strip()
-    if not text:
-        return []
-    tokens = text.split()
-    return tokens
-
-
-class FullTokenizer(object):
-    """Runs end-to-end tokenziation."""
-    def __init__(self, vocab_file, do_lower_case=True):
-        self.vocab = load_vocab(vocab_file)
-        self.inv_vocab = {v: k for k, v in self.vocab.items()}
-        self.basic_tokenizer = BasicTokenizer(do_lower_case=do_lower_case)
-        self.wordpiece_tokenizer = WordpieceTokenizer(vocab=self.vocab)
-
-    def tokenize(self, text):
-        split_tokens = []
-        for token in self.basic_tokenizer.tokenize(text):
-            for sub_token in self.wordpiece_tokenizer.tokenize(token):
-                split_tokens.append(sub_token)
-
-        return split_tokens
-
-    def convert_tokens_to_ids(self, tokens):
-        return convert_by_vocab(self.vocab, tokens)
-
-    def convert_ids_to_tokens(self, ids):
-        return convert_by_vocab(self.inv_vocab, ids)
-
-
-class BasicTokenizer(object):
-    """Runs basic tokenization (punctuation splitting, lower casing, etc.)."""
-    def __init__(self, do_lower_case=True):
-        """Constructs a BasicTokenizer.
-
-    Args:
-      do_lower_case: Whether to lower case the input.
-    """
-        self.do_lower_case = do_lower_case
-
-    def tokenize(self, text):
-        """Tokenizes a piece of text."""
-        text = convert_to_unicode(text)
-        text = self._clean_text(text)
-
-        # This was added on November 1st, 2018 for the multilingual and Chinese
-        # models. This is also applied to the English models now, but it doesn't
-        # matter since the English models were not trained on any Chinese data
-        # and generally don't have any Chinese data in them (there are Chinese
-        # characters in the vocabulary because Wikipedia does have some Chinese
-        # words in the English Wikipedia.).
-        text = self._tokenize_chinese_chars(text)
-
-        orig_tokens = whitespace_tokenize(text)
-        split_tokens = []
-        for token in orig_tokens:
-            if self.do_lower_case:
-                token = token.lower()
-                token = self._run_strip_accents(token)
-            split_tokens.extend(self._run_split_on_punc(token))
-
-        output_tokens = whitespace_tokenize(" ".join(split_tokens))
-        return output_tokens
-
-    def _run_strip_accents(self, text):
-        """Strips accents from a piece of text."""
-        text = unicodedata.normalize("NFD", text)
-        output = []
-        for char in text:
-            cat = unicodedata.category(char)
-            if cat == "Mn":
-                continue
-            output.append(char)
-        return "".join(output)
-
-    def _run_split_on_punc(self, text):
-        """Splits punctuation on a piece of text."""
-        chars = list(text)
-        i = 0
-        start_new_word = True
-        output = []
-        while i < len(chars):
-            char = chars[i]
-            if _is_punctuation(char):
-                output.append([char])
-                start_new_word = True
-            else:
-                if start_new_word:
-                    output.append([])
-                start_new_word = False
-                output[-1].append(char)
-            i += 1
-
-        return ["".join(x) for x in output]
-
-    def _tokenize_chinese_chars(self, text):
-        """Adds whitespace around any CJK character."""
-        output = []
-        for char in text:
-            cp = ord(char)
-            if self._is_chinese_char(cp):
-                output.append(" ")
-                output.append(char)
-                output.append(" ")
-            else:
-                output.append(char)
-        return "".join(output)
-
-    def _is_chinese_char(self, cp):
-        """Checks whether CP is the codepoint of a CJK character."""
-        # This defines a "chinese character" as anything in the CJK Unicode block:
-        #   https://en.wikipedia.org/wiki/CJK_Unified_Ideographs_(Unicode_block)
-        #
-        # Note that the CJK Unicode block is NOT all Japanese and Korean characters,
-        # despite its name. The modern Korean Hangul alphabet is a different block,
-        # as is Japanese Hiragana and Katakana. Those alphabets are used to write
-        # space-separated words, so they are not treated specially and handled
-        # like the all of the other languages.
-        if ((cp >= 0x4E00 and cp <= 0x9FFF) or  #
-            (cp >= 0x3400 and cp <= 0x4DBF) or  #
-            (cp >= 0x20000 and cp <= 0x2A6DF) or  #
-            (cp >= 0x2A700 and cp <= 0x2B73F) or  #
-            (cp >= 0x2B740 and cp <= 0x2B81F) or  #
-            (cp >= 0x2B820 and cp <= 0x2CEAF) or
-            (cp >= 0xF900 and cp <= 0xFAFF) or  #
-            (cp >= 0x2F800 and cp <= 0x2FA1F)):  #
-            return True
-
-        return False
-
-    def _clean_text(self, text):
-        """Performs invalid character removal and whitespace cleanup on text."""
-        output = []
-        for char in text:
-            cp = ord(char)
-            if cp == 0 or cp == 0xfffd or _is_control(char):
-                continue
-            if _is_whitespace(char):
-                output.append(" ")
-            else:
-                output.append(char)
-        return "".join(output)
-
-
-class WordpieceTokenizer(object):
-    """Runs WordPiece tokenziation."""
-    def __init__(self, vocab, unk_token="[UNK]", max_input_chars_per_word=200):
-        self.vocab = vocab
-        self.unk_token = unk_token
-        self.max_input_chars_per_word = max_input_chars_per_word
-
-    def tokenize(self, text):
-        """Tokenizes a piece of text into its word pieces.
-
-    This uses a greedy longest-match-first algorithm to perform tokenization
-    using the given vocabulary.
-
-    For example:
-      input = "unaffable"
-      output = ["un", "##aff", "##able"]
-
-    Args:
-      text: A single token or whitespace separated tokens. This should have
-        already been passed through `BasicTokenizer.
-
-    Returns:
-      A list of wordpiece tokens.
-    """
-
-        text = convert_to_unicode(text)
-
-        output_tokens = []
-        for token in whitespace_tokenize(text):
-            chars = list(token)
-            if len(chars) > self.max_input_chars_per_word:
-                output_tokens.append(self.unk_token)
-                continue
-
-            is_bad = False
-            start = 0
-            sub_tokens = []
-            while start < len(chars):
-                end = len(chars)
-                cur_substr = None
-                while start < end:
-                    substr = "".join(chars[start:end])
-                    if start > 0:
-                        substr = "##" + substr
-                    if substr in self.vocab:
-                        cur_substr = substr
-                        break
-                    end -= 1
-                if cur_substr is None:
-                    is_bad = True
-                    break
-                sub_tokens.append(cur_substr)
-                start = end
-
-            if is_bad:
-                output_tokens.append(self.unk_token)
-            else:
-                output_tokens.extend(sub_tokens)
-        return output_tokens
-
-
-def _is_whitespace(char):
-    """Checks whether `chars` is a whitespace character."""
-    # \t, \n, and \r are technically contorl characters but we treat them
-    # as whitespace since they are generally considered as such.
-    if char == " " or char == "\t" or char == "\n" or char == "\r":
-        return True
-    cat = unicodedata.category(char)
-    if cat == "Zs":
-        return True
-    return False
-
-
-def _is_control(char):
-    """Checks whether `chars` is a control character."""
-    # These are technically control characters but we count them as whitespace
-    # characters.
-    if char == "\t" or char == "\n" or char == "\r":
-        return False
-    cat = unicodedata.category(char)
-    if cat in ("Cc", "Cf"):
-        return True
-    return False
-
-
-def _is_punctuation(char):
-    """Checks whether `chars` is a punctuation character."""
-    cp = ord(char)
-    # We treat all non-letter/number ASCII as punctuation.
-    # Characters such as "^", "$", and "`" are not in the Unicode
-    # Punctuation class but we treat them as punctuation anyways, for
-    # consistency.
-    if ((cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64)
-            or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126)):
-        return True
-    cat = unicodedata.category(char)
-    if cat.startswith("P"):
-        return True
-    return False