File size: 4,447 Bytes
2d8da09
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
# Copyright (c) 2022, NVIDIA CORPORATION & AFFILIATES.  All rights reserved.
#
# 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.

import argparse
import json
import re
import string
from collections import Counter
from rouge_score import rouge_scorer


"""
This script can be used to calcualte exact match and F1 scores for many different tasks. 
The file "squad_test_predictions.jsonl" is assumed to be generated by the 
`examples/nlp/language_modeling/tuning/megatron_gpt_peft_eval.py` script

Example command for GPT Preds

    ```
    python peft_metric_calc.py \
        --pred_file squad_test_predictions.jsonl \
        --label_field "original_answers" \
    ```


"""


def normalize_answer(s):
    """Lower text and remove punctuation, articles and extra whitespace."""

    def remove_articles(text):
        return re.sub(r'\b(a|an|the)\b', ' ', text)

    def white_space_fix(text):
        return ' '.join(text.split())

    def remove_punc(text):
        exclude = set(string.punctuation)
        return ''.join(ch for ch in text if ch not in exclude)

    def lower(text):
        return text.lower()

    return white_space_fix(remove_articles(remove_punc(lower(s))))


def f1_score(prediction, ground_truth):
    prediction_tokens = normalize_answer(prediction).split()
    ground_truth_tokens = normalize_answer(ground_truth).split()
    common = Counter(prediction_tokens) & Counter(ground_truth_tokens)
    num_same = sum(common.values())
    if num_same == 0:
        return 0
    precision = 1.0 * num_same / len(prediction_tokens)
    recall = 1.0 * num_same / len(ground_truth_tokens)
    f1 = (2 * precision * recall) / (precision + recall)
    return f1


def exact_match_score(prediction, ground_truth):
    return normalize_answer(prediction) == normalize_answer(ground_truth)


def metric_max_over_ground_truths(metric_fn, prediction, ground_truths):
    scores_for_ground_truths = []
    for ground_truth in ground_truths:
        score = metric_fn(prediction, ground_truth)
        scores_for_ground_truths.append(score)
    return max(scores_for_ground_truths)


def main():
    parser = argparse.ArgumentParser(description='Process some integers.')
    parser.add_argument(
        '--pred_file',
        type=str,
        help="Text file with test set prompts + model predictions. Prediction file can be made by running NeMo/examples/nlp/language_modeling/megatron_gpt_prompt_learning_eval.py",
    )
    parser.add_argument(
        '--pred_field',
        type=str,
        help="The field in the json file that contains the prediction tokens",
        default="pred",
    )
    parser.add_argument(
        '--label_field',
        type=str,
        help="The field in the json file that contains the ground truth tokens",
        default="label",
    )

    args = parser.parse_args()

    pred_file = args.pred_file
    scorer = rouge_scorer.RougeScorer(['rougeL'], use_stemmer=True)
    preds = open(pred_file, encoding="utf-8").readlines()
    f1 = exact_match = total = r_score = 0

    for i in range(len(preds)):
        pred_line = json.loads(preds[i])

        pred_answer = pred_line[args.pred_field]
        true_answers = pred_line[args.label_field]
        if not isinstance(true_answers, list):
            true_answers = [true_answers]

        r_scores = []
        for ta in true_answers:
            r_scores.append(scorer.score(ta, pred_answer)['rougeL'].fmeasure)
        r_score += max(r_scores)
        exact_match += metric_max_over_ground_truths(exact_match_score, pred_answer, true_answers)
        f1 += metric_max_over_ground_truths(f1_score, pred_answer, true_answers)
        total += 1

    exact_match = 100.0 * exact_match / total
    f1 = 100.0 * f1 / total
    r_score = 100 * (r_score / total)
    res = {'exact_match': exact_match, 'f1': f1, "rougeL": r_score, 'total': total}
    print('\t'.join([f"{k} {v:.3f}" for k, v in res.items()]))


if __name__ == "__main__":
    main()