-
Notifications
You must be signed in to change notification settings - Fork 20
/
gpt4_eval.py
79 lines (69 loc) · 3.53 KB
/
gpt4_eval.py
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
# Evaluate on all datasets in VQAScore paper
# python gpt4_eval.py --model gpt-4-turbo
# python gpt4_eval.py --model gpt-4o
import argparse
import os
import t2v_metrics
from dataset import Winoground, EqBen_Mini, StanfordT23D, TIFA160_DSG, Flickr8K_CF, SeeTrue, Pickapic_v1, T2VScore
def config():
parser = argparse.ArgumentParser()
parser.add_argument("--root_dir", default="./datasets", type=str,
help='Root directory for saving datasets.')
parser.add_argument("--cache_dir", default=t2v_metrics.constants.HF_CACHE_DIR, type=str)
parser.add_argument("--device", default="cuda", type=str)
parser.add_argument("--batch_size", default=16, type=int)
parser.add_argument("--model", default="gpt-4-turbo", type=str)
parser.add_argument("--question", default=None, type=str)
parser.add_argument("--answer", default=None, type=str)
parser.add_argument("--openai_key", default=None, type=str)
parser.add_argument("--openai_key_path", default='./_OPENAI_API_KEY.txt', type=str)
parser.add_argument("--top_logprobs", type=int, default=20)
parser.add_argument("--detail", type=str, default='auto', choices=['low', 'auto', 'high'])
return parser.parse_args()
def main():
args = config()
if not os.path.exists(args.root_dir):
os.makedirs(args.root_dir)
assert not (args.openai_key is None and args.openai_key_path is None), "Please provide either openai_key or openai_key_path."
if args.openai_key is None:
args.openai_key = open(args.openai_key_path, 'r').read().strip()
score_func = t2v_metrics.get_score_model(
model=args.model, device=args.device, cache_dir=args.cache_dir, openai_key=args.openai_key, top_logprobs=args.top_logprobs)
kwargs = {}
if args.question is not None:
print(f"Using question template: {args.question}")
kwargs['question_template'] = args.question
if args.answer is not None:
print(f"Using answer template: {args.answer}")
kwargs['answer_template'] = args.answer
print(f"Performance of {args.model}.")
for dataset_cls in [
Winoground,
# EqBen_Mini,
# TIFA160_DSG,
# Pickapic_v1,
# SeeTrue,
# StanfordT23D,
# T2VScore,
# Flickr8K_CF,
]:
print(f"Evaluating on {dataset_cls.__name__}.")
dataset = dataset_cls(root_dir=args.root_dir)
# check file size under 15mb
for item in dataset:
images = item['images']
for image in images:
assert os.path.getsize(image) < 15 * 1024 * 1024, f"File size of {image} is {os.path.getsize(image)/1048576} bytes, which is larger than 15mb."
img_type = image.split('.')[-1]
assert img_type in ['png', 'jpeg', 'jpg', 'gif', 'webp'], f"Image type {img_type} is not supported."
scores = score_func.batch_forward(dataset, batch_size=args.batch_size, **kwargs).cpu()
score_path = f'./{args.model}_{args.detail}_{dataset_cls.__name__}.pt'
# score_path = f'./{args.model}_{args.detail}_{dataset_cls.__name__}_no.pt' # when answer is "No"
# score_path = f'./{args.model}_{args.detail}_{dataset_cls.__name__}_does_this_image_show_yes_or_no.pt'
# score_path = f'./{args.model}_{args.detail}_{dataset_cls.__name__}_yes_or_no.pt'
# score_path = f'./{args.model}_{args.detail}_{dataset_cls.__name__}_does_this_figure_show.pt'
import torch
torch.save(scores, score_path)
dataset.evaluate_scores(scores)
if __name__ == "__main__":
main()