File size: 8,279 Bytes
3334467 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 | from enum import Enum
import numpy as np
import torch
import torch.distributed as dist
IGNORE_INDEX = -100
IMAGE_TOKEN_INDEX = -200
DEFAULT_IMAGE_TOKEN = "<image>"
DEFAULT_IMAGE_PATCH_TOKEN = "<im_patch>"
DEFAULT_IM_START_TOKEN = "<im_start>"
DEFAULT_IM_END_TOKEN = "<im_end>"
SHORT_QUESTION_LIST = [
DEFAULT_IMAGE_TOKEN + "\n" + "Can you segment the {class_name} in this image?",
DEFAULT_IMAGE_TOKEN + "\n" + "Please segment the {class_name} in this image.",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "What is {class_name} in this image? Please respond with segmentation mask.",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "What is {class_name} in this image? Please output segmentation mask.",
]
LONG_QUESTION_LIST = [
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Please respond with segmentation mask.",
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Please output segmentation mask.",
]
EXPLANATORY_QUESTION_LIST = [
"Please output segmentation mask and explain why.",
"Please output segmentation mask and explain the reason.",
"Please output segmentation mask and give some explanation.",
]
ANSWER_LIST = [
"It is [SEG].",
"Sure, [SEG].",
"Sure, it is [SEG].",
"Sure, the segmentation result is [SEG].",
"[SEG].",
]
SINGLE_ANSWER_LIST = [
"{class_name} is [SEG].",
"The segmentation result of {class_name} is [SEG].",
"[SEG]."
]
MULTI_ANSWER_LIST = [
"{class_name} are {seg}, separately.",
"{class_name} are {seg}.",
"Sure, {class_name} are {seg}, separately.",
"Sure, {class_name} are {seg}.",
"the segmentation result of {class_name} are {seg}.",
"the segmentation result of {class_name} are {seg}, separately.",
"Sure, the segmentation result of {class_name} are {seg}.",
"Sure, the segmentation result of {class_name} are {seg}, separately.",
"Sure, they are {seg}.",
"They are {seg}.",
"{seg}."
]
MR_SINGLE_ANSWER_LIST = [
"{class_name} is [SEG].",
]
MR_MULTI_ANSWER_LIST = [
"{class_name} are {seg}, separately.",
"{class_name} are {seg}.",
"Sure, {class_name} are {seg}, separately.",
"Sure, {class_name} are {seg}.",
"the segmentation result of {class_name} are {seg}.",
"the segmentation result of {class_name} are {seg}, separately.",
"Sure, the segmentation result of {class_name} are {seg}.",
"Sure, the segmentation result of {class_name} are {seg}, separately.",
]
EXPAND_LONG_QUESTION_LIST = [
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Provide the segmentation mask.",
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Output the segmentation mask.",
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Please show the segmentation mask.",
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} I'd appreciate segmentation masks.",
DEFAULT_IMAGE_TOKEN + "\n" + "{sent} Please highlight the segmentation mask.",
]
EXPAND_QUESTION_LIST = [
DEFAULT_IMAGE_TOKEN + "\n" + "Could you identify the {class_name} in this picture?",
DEFAULT_IMAGE_TOKEN + "\n" + "Are you able to delineate the {class_name} in the image?",
DEFAULT_IMAGE_TOKEN + "\n" + "Can you pinpoint the {class_name} in this photo?",
DEFAULT_IMAGE_TOKEN + "\n" + "Is it possible for you to highlight the {class_name} in this image?",
DEFAULT_IMAGE_TOKEN + "\n" + "Can you discern the {class_name} in the given picture?",
DEFAULT_IMAGE_TOKEN + "\n" + "Can you provide me with asegment of the {class_name}?",
DEFAULT_IMAGE_TOKEN + "\n" + "Please perform image segmentation to isolate the {class_name} in this image.",
DEFAULT_IMAGE_TOKEN + "\n" + "Help me segment the {class_name}.",
DEFAULT_IMAGE_TOKEN + "\n" + "Would you be willing to segment the {class_name}?",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "Can you identify {class_name} in this picture? Please provide a segmentation mask.",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "Could you point out {class_name} in this image and show it with a segmentation mask?",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "In this image, where is {class_name}? I'd appreciate a segmentation mask.",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "Please highlight {class_name} in this image using a segmentation mask.",
DEFAULT_IMAGE_TOKEN
+ "\n"
+ "In the picture provided, can you show where {class_name} is with a segmentation mask?",
]
class Summary(Enum):
NONE = 0
AVERAGE = 1
SUM = 2
COUNT = 3
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self, name, fmt=":f", summary_type=Summary.AVERAGE):
self.name = name
self.fmt = fmt
self.summary_type = summary_type
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
def all_reduce(self):
device = "cuda" if torch.cuda.is_available() else "cpu"
if isinstance(self.sum, np.ndarray):
total = torch.tensor(
self.sum.tolist()
+ [
self.count,
],
dtype=torch.float32,
device=device,
)
else:
total = torch.tensor(
[self.sum, self.count], dtype=torch.float32, device=device
)
dist.all_reduce(total, dist.ReduceOp.SUM, async_op=False)
if total.shape[0] > 2:
self.sum, self.count = total[:-1].cpu().numpy(), total[-1].cpu().item()
else:
self.sum, self.count = total.tolist()
self.avg = self.sum / (self.count + 1e-5)
def __str__(self):
fmtstr = "{name} {val" + self.fmt + "} ({avg" + self.fmt + "})"
return fmtstr.format(**self.__dict__)
def summary(self):
fmtstr = ""
if self.summary_type is Summary.NONE:
fmtstr = ""
elif self.summary_type is Summary.AVERAGE:
fmtstr = "{name} {avg:.3f}"
elif self.summary_type is Summary.SUM:
fmtstr = "{name} {sum:.3f}"
elif self.summary_type is Summary.COUNT:
fmtstr = "{name} {count:.3f}"
else:
raise ValueError("invalid summary type %r" % self.summary_type)
return fmtstr.format(**self.__dict__)
def intersectionAndUnionGPU(output, target, K, ignore_index=255):
assert output.dim() in [1, 2, 3]
assert output.shape == target.shape
output = output.view(-1)
target = target.view(-1)
output[target == ignore_index] = ignore_index
intersection = output[output == target]
area_intersection = torch.histc(intersection, bins=K, min=0, max=K - 1)
area_output = torch.histc(output, bins=K, min=0, max=K - 1)
area_target = torch.histc(target, bins=K, min=0, max=K - 1)
area_union = area_output + area_target - area_intersection
return area_intersection, area_union, area_target
class ProgressMeter(object):
def __init__(self, num_batches, meters, prefix=""):
self.batch_fmtstr = self._get_batch_fmtstr(num_batches)
self.meters = meters
self.prefix = prefix
def display(self, batch):
entries = [self.prefix + self.batch_fmtstr.format(batch)]
entries += [str(meter) for meter in self.meters]
print("\t".join(entries))
def display_summary(self):
entries = [" *"]
entries += [meter.summary() for meter in self.meters]
print(" ".join(entries))
def _get_batch_fmtstr(self, num_batches):
num_digits = len(str(num_batches // 1))
fmt = "{:" + str(num_digits) + "d}"
return "[" + fmt + "/" + fmt.format(num_batches) + "]"
def dict_to_cuda(input_dict):
for k, v in input_dict.items():
if isinstance(input_dict[k], torch.Tensor):
input_dict[k] = v.cuda(non_blocking=True)
elif (
isinstance(input_dict[k], list)
and len(input_dict[k]) > 0
and isinstance(input_dict[k][0], torch.Tensor)
):
input_dict[k] = [ele.cuda(non_blocking=True) for ele in v]
return input_dict
|