-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathinference_mmu_caption.py
136 lines (108 loc) · 5.98 KB
/
inference_mmu_caption.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
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
import os
os.environ["TOKENIZERS_PARALLELISM"] = "true"
from PIL import Image
from tqdm import tqdm
import numpy as np
import torch
import wandb
from models import Showo, MAGVITv2, CLIPVisionTower
from training.prompting_utils import UniversalPrompting, create_attention_mask_for_mmu, create_attention_mask_for_mmu_vit
from training.utils import get_config, flatten_omega_conf, image_transform
from transformers import AutoTokenizer
from transformers import CLIPImageProcessor
import json
from llava.llava import conversation as conversation_lib
conversation_lib.default_conversation = conversation_lib.conv_templates["phi1.5"]
SYSTEM_PROMPT = "A chat between a curious user and an artificial intelligence assistant. " \
"The assistant gives helpful, detailed, and polite answers to the user's questions."
SYSTEM_PROMPT_LEN = 28
def get_vq_model_class(model_type):
if model_type == "magvitv2":
return MAGVITv2
else:
raise ValueError(f"model_type {model_type} not supported.")
if __name__ == '__main__':
config = get_config()
resume_wandb_run = config.wandb.resume
run_id = config.wandb.get("run_id", None)
if run_id is None:
resume_wandb_run = False
run_id = wandb.util.generate_id()
config.wandb.run_id = run_id
wandb_config = {k: v for k, v in flatten_omega_conf(config, resolve=True)}
wandb.init(
project="demo",
name=config.experiment.name + '_mmu',
config=wandb_config,
)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
tokenizer = AutoTokenizer.from_pretrained(config.model.showo.llm_model_path, padding_side="left")
uni_prompting = UniversalPrompting(tokenizer, max_text_len=config.dataset.preprocessing.max_seq_length,
special_tokens=("<|soi|>", "<|eoi|>", "<|sov|>", "<|eov|>", "<|t2i|>", "<|mmu|>", "<|t2v|>", "<|v2v|>", "<|lvg|>"),
ignore_id=-100, cond_dropout_prob=config.training.cond_dropout_prob)
vq_model = get_vq_model_class(config.model.vq_model.type)
vq_model = vq_model.from_pretrained(config.model.vq_model.vq_model_name).to(device)
vq_model.requires_grad_(False)
vq_model.eval()
vision_tower_name = "openai/clip-vit-large-patch14-336"
vision_tower = CLIPVisionTower(vision_tower_name).to(device)
clip_image_processor = CLIPImageProcessor.from_pretrained(vision_tower_name)
model = Showo.from_pretrained(config.model.showo.pretrained_model_path).to(device)
model.eval()
temperature = 0.8 # 1.0 = no change, < 1.0 = less random, > 1.0 = more random, in predictions
top_k = 1 # retain only the top_k most likely tokens, clamp others to have 0 probability
with open('datasets/journeydb/initial_data.json', "r") as f:
data = json.load(f)
results = []
caption = []
for i in tqdm(range(len(data))):
image_path = data[i]["img_path"]
image_ori = Image.open(image_path).convert("RGB")
image = image_transform(image_ori, resolution=config.dataset.params.resolution).to(device)
image = image.unsqueeze(0)
pixel_values = clip_image_processor.preprocess(image_ori, return_tensors="pt")["pixel_values"][0]
image_tokens = vq_model.get_code(image) + len(uni_prompting.text_tokenizer)
batch_size = 1
question = "Give a caption for this image."
input_ids = uni_prompting.text_tokenizer(['USER: \n' + question + ' ASSISTANT:'])[
'input_ids']
input_ids = torch.tensor(input_ids).to(device)
input_ids = torch.cat([
(torch.ones(input_ids.shape[0], 1) * uni_prompting.sptids_dict['<|mmu|>']).to(device),
(torch.ones(input_ids.shape[0], 1) * uni_prompting.sptids_dict['<|soi|>']).to(device),
image_tokens,
(torch.ones(input_ids.shape[0], 1) * uni_prompting.sptids_dict['<|eoi|>']).to(device),
(torch.ones(input_ids.shape[0], 1) * uni_prompting.sptids_dict['<|sot|>']).to(device),
input_ids
], dim=1).long()
attention_mask = create_attention_mask_for_mmu(input_ids.to(device),
eoi_id=int(uni_prompting.sptids_dict['<|eoi|>']))
for top_k in [1, 2, 3, 4]:
if top_k != 1:
for tempterature in [0.1, 0.7, 2.0]:
cont_toks_list = model.mmu_generate(input_ids, attention_mask=attention_mask,
max_new_tokens=config.max_new_tokens, top_k=top_k,
temperature=tempterature,
eot_token=uni_prompting.sptids_dict['<|eot|>'])
cont_toks_list = torch.stack(cont_toks_list).squeeze()[None]
text = uni_prompting.text_tokenizer.batch_decode(cont_toks_list, skip_special_tokens=True)
caption.append(text[0])
else:
tempterature = 1
cont_toks_list = model.mmu_generate(input_ids, attention_mask=attention_mask,
max_new_tokens=config.max_new_tokens, top_k=top_k,
temperature=tempterature,
eot_token=uni_prompting.sptids_dict['<|eot|>'])
cont_toks_list = torch.stack(cont_toks_list).squeeze()[None]
text = uni_prompting.text_tokenizer.batch_decode(cont_toks_list, skip_special_tokens=True)
caption.append(text[0])
with open(f"datasets/journeydb/understanding_caption_results", "a") as f:
json.dump({
"id": data[i]["id"],
"img_path": data[i]["img_path"],
"prompt": data[i]["prompt"],
"caption": data[i]["caption"],
"caption_mmu": caption
}, f, ensure_ascii=False, indent=4)
f.write(',\n')
caption = []