UGround / llava /mm_utils.py
BoyuNLP's picture
init
3bbba47
from PIL import Image
from io import BytesIO
import base64
import torch
import math
import ast
from transformers import StoppingCriteria
from llava.constants import IMAGE_TOKEN_INDEX
def select_best_resolution(original_size, possible_resolutions):
"""
Selects the best resolution from a list of possible resolutions based on the original size.
Args:
original_size (tuple): The original size of the image in the format (width, height).
possible_resolutions (list): A list of possible resolutions in the format [(width1, height1), (width2, height2), ...].
Returns:
tuple: The best fit resolution in the format (width, height).
"""
original_width, original_height = original_size
best_fit = None
max_effective_resolution = 0
min_wasted_resolution = float('inf')
for width, height in possible_resolutions:
scale = min(width / original_width, height / original_height)
downscaled_width, downscaled_height = int(original_width * scale), int(original_height * scale)
effective_resolution = min(downscaled_width * downscaled_height, original_width * original_height)
wasted_resolution = (width * height) - effective_resolution
if effective_resolution > max_effective_resolution or (effective_resolution == max_effective_resolution and wasted_resolution < min_wasted_resolution):
max_effective_resolution = effective_resolution
min_wasted_resolution = wasted_resolution
best_fit = (width, height)
return best_fit
def get_resized_ui_resolution(original_size):
"""
Selects the best resolution from a list of possible resolutions based on the original size.
Args:
original_size (tuple): The original size of the image in the format (width, height).
possible_resolutions (list): A list of possible resolutions in the format [(width1, height1), (width2, height2), ...].
Returns:
tuple: The best fit resolution in the format (width, height).
"""
original_width, original_height = original_size
# best_fit = None
new_width = nearest_multiple_of_224_at_least_224(original_width,upperbound=26880)
scale_factor = new_width / original_width
new_height_unpadded = int(original_height * scale_factor)
new_height_padded = nearest_multiple_of_224_at_least_224(new_height_unpadded,ceiling=True)
best_fit=(new_width,new_height_padded)
return best_fit
def resize_and_pad_image(image, target_resolution):
"""
Resize and pad an image to a target resolution while maintaining aspect ratio.
Args:
image (PIL.Image.Image): The input image.
target_resolution (tuple): The target resolution (width, height) of the image.
Returns:
PIL.Image.Image: The resized and padded image.
"""
original_width, original_height = image.size
target_width, target_height = target_resolution
scale_w = target_width / original_width
scale_h = target_height / original_height
if scale_w < scale_h:
new_width = target_width
new_height = min(math.ceil(original_height * scale_w), target_height)
else:
new_height = target_height
new_width = min(math.ceil(original_width * scale_h), target_width)
# Resize the image
resized_image = image.resize((new_width, new_height))
new_image = Image.new('RGB', (target_width, target_height), (0, 0, 0))
paste_x = (target_width - new_width) // 2
paste_y = (target_height - new_height) // 2
new_image.paste(resized_image, (paste_x, paste_y))
return new_image
def resize_and_pad_ui_image(img):
"""
Resize and pad an image to a target resolution while maintaining aspect ratio.
Args:
img (PIL.Image.Image): The input image.
Returns:
PIL.Image.Image: The resized and padded image.
"""
#TODO: [USE THIS ->->] True training
orig_width, orig_height = img.size
# print("DEBUG--- (orig_width, orig_height)", (orig_width, orig_height))
# target_width, target_height = target_resolution
# scale_w = target_width / original_width
# scale_h = target_height / original_height
new_width = nearest_multiple_of_224_at_least_224(orig_width, upperbound=26880)
# print("new_width",new_width)
scale_factor = new_width / orig_width
new_height_unpadded = min(int(orig_height * scale_factor),26880)
# print("new_height_unpadded", new_height_unpadded)
img_resized = img.resize((new_width, new_height_unpadded))
new_height_padded = nearest_multiple_of_224_at_least_224(new_height_unpadded,ceiling=True,upperbound=268800)
# print("new_height_padded", new_height_padded)
img_padded = Image.new('RGB', (new_width, new_height_padded), (0, 0, 0))
img_padded.paste(img_resized, (0, 0))
new_size=(new_width,new_height_padded)
#TODO [DO NOT USE THIS!!!!] TEST FOR UPPERBOUND square
# orig_width, orig_height = img.size
#
# # target_width, target_height = target_resolution
#
# # scale_w = target_width / original_width
# # scale_h = target_height / original_height
#
# new_width = 1344
#
# # print("new_width",new_width)
#
# scale_factor = new_width / orig_width
#
# new_height_unpadded = 1344
# # print("new_height_unpadded", new_height_unpadded)
#
# img_resized = img.resize((new_width, new_height_unpadded))
#
# new_height_padded = nearest_multiple_of_224_at_least_224(new_height_unpadded, ceiling=True, upperbound=1344)
# # print("new_height_padded", new_height_padded)
#
# img_padded = Image.new('RGB', (new_width, new_height_padded), (0, 0, 0))
# img_padded.paste(img_resized, (0, 0))
#
# new_size = (new_width, new_height_padded)
# TODO [DO NOT USE THIS!!!!] TEST FOR UPPERBOUND square
#
# if scale_w < scale_h:
# new_width = target_width
# new_height = min(math.ceil(original_height * scale_w), target_height)
# else:
# new_height = target_height
# new_width = min(math.ceil(original_width * scale_h), target_width)
# Resize the image
# resized_image = image.resize((new_width, new_height))
# new_image = Image.new('RGB', (target_width, target_height), (0, 0, 0))
# paste_x = (target_width - new_width) // 2
# paste_y = (target_height - new_height) // 2
# new_image.paste(resized_image, (paste_x, paste_y))
return img_padded,new_size
def resize_and_pad_image_to_top_left(image, target_resolution):
"""
Resize and pad an image to a target resolution while maintaining aspect ratio,
and align the image to the top-left corner of the new image.
Args:
image (PIL.Image.Image): The input image.
target_resolution (tuple): The target resolution (width, height) of the image.
Returns:
PIL.Image.Image: The resized and padded image aligned to the top-left corner.
"""
original_width, original_height = image.size
target_width, target_height = target_resolution
scale_w = target_width / original_width
scale_h = target_height / original_height
if scale_w < scale_h:
new_width = target_width
new_height = min(math.ceil(original_height * scale_w), target_height)
else:
new_height = target_height
new_width = min(math.ceil(original_width * scale_h), target_width)
# Resize the image
resized_image = image.resize((new_width, new_height))
# Create a new image with a black background
new_image = Image.new('RGB', (target_width, target_height), (0, 0, 0))
# Paste the resized image into the new image, aligned to the top-left corner
paste_x = 0 # Align to the left
paste_y = 0 # Align to the top
new_image.paste(resized_image, (paste_x, paste_y))
return new_image
def divide_to_patches(image, patch_size):
"""
Divides an image into patches of a specified size.
Args:
image (PIL.Image.Image): The input image.
patch_size (int): The size of each patch.
Returns:
list: A list of PIL.Image.Image objects representing the patches.
"""
patches = []
width, height = image.size
for i in range(0, height, patch_size):
for j in range(0, width, patch_size):
box = (j, i, j + patch_size, i + patch_size)
patch = image.crop(box)
patches.append(patch)
return patches
def get_anyres_image_grid_shape(image_size, grid_pinpoints, patch_size):
"""
Calculate the shape of the image patch grid after the preprocessing for images of any resolution.
Args:
image_size (tuple): The size of the input image in the format (width, height).
grid_pinpoints (str): A string representation of a list of possible resolutions.
patch_size (int): The size of each image patch.
Returns:
tuple: The shape of the image patch grid in the format (width, height).
"""
if type(grid_pinpoints) is list:
possible_resolutions = grid_pinpoints
else:
possible_resolutions = ast.literal_eval(grid_pinpoints)
width, height = select_best_resolution(image_size, possible_resolutions)
return width // patch_size, height // patch_size
def get_anyres_image_grid_shape_ui(image_size, patch_size):
"""
Calculate the shape of the image patch grid after the preprocessing for images of any resolution.
Args:
image_size (tuple): The size of the input image in the format (width, height).
grid_pinpoints (str): A string representation of a list of possible resolutions.
patch_size (int): The size of each image patch.
Returns:
tuple: The shape of the image patch grid in the format (width, height).
"""
# if type(grid_pinpoints) is list:
# possible_resolutions = grid_pinpoints
# else:
# possible_resolutions = ast.literal_eval(grid_pinpoints)
width, height = image_size
# width, height = get_resized_ui_resolution(image_size)
return width // patch_size, height // patch_size
def process_anyres_image(image, processor, grid_pinpoints):
"""
Process an image with variable resolutions.
Args:
image (PIL.Image.Image): The input image to be processed.
processor: The image processor object.
grid_pinpoints (str): A string representation of a list of possible resolutions.
Returns:
torch.Tensor: A tensor containing the processed image patches.
"""
if type(grid_pinpoints) is list:
possible_resolutions = grid_pinpoints
else:
possible_resolutions = ast.literal_eval(grid_pinpoints)
best_resolution = select_best_resolution(image.size, possible_resolutions)
image_padded = resize_and_pad_image(image, best_resolution)
patches = divide_to_patches(image_padded, processor.crop_size['height'])
image_original_resize = image.resize((processor.size['shortest_edge'], processor.size['shortest_edge']))
image_patches = [image_original_resize] + patches
image_patches = [processor.preprocess(image_patch, return_tensors='pt')['pixel_values'][0]
for image_patch in image_patches]
return torch.stack(image_patches, dim=0)
def nearest_multiple_of_224_at_least_224(num,ceiling=False,upperbound=26880):
if num <= 224:
return 224
division, remainder = divmod(num, 224)
if ceiling and remainder>0:
return (division + 1) * 224
if remainder < 112:
return min(division * 224,upperbound)
else:
return min((division + 1) * 224,upperbound)
def process_anyres_ui_image(image, processor,fusion=False):
"""
Process an image with variable resolutions.
Args:
image (PIL.Image.Image): The input image to be processed.
processor: The image processor object.
grid_pinpoints (str): A string representation of a list of possible resolutions.
Returns:
torch.Tensor: A tensor containing the processed image patches.
"""
# if type(grid_pinpoints) is list:
# possible_resolutions = grid_pinpoints
# else:
# # possible_resolutions = ast.literal_eval(grid_pinpoints)
# best_resolution = select_best_resolution(image.size, possible_resolutions)
image_padded,new_size = resize_and_pad_ui_image(image)
patches = divide_to_patches(image_padded, 224)
if fusion:
image_original_resize = image.resize((224, 224))
image_patches = [image_original_resize] + patches
else:
image_patches = patches
# if len(image_patches)==2:
# print(f"\n len image_patches: {len(image_patches)}")
image_patches = [processor.preprocess(image_patch, return_tensors='pt')['pixel_values'][0]
for image_patch in image_patches]
return torch.stack(image_patches, dim=0),new_size
def load_image_from_base64(image):
return Image.open(BytesIO(base64.b64decode(image)))
def expand2square(pil_img, background_color):
width, height = pil_img.size
if width == height:
return pil_img
elif width > height:
result = Image.new(pil_img.mode, (width, width), background_color)
result.paste(pil_img, (0, (width - height) // 2))
return result
else:
result = Image.new(pil_img.mode, (height, height), background_color)
result.paste(pil_img, ((height - width) // 2, 0))
return result
def process_images(images, image_processor, model_cfg):
image_aspect_ratio = getattr(model_cfg, "image_aspect_ratio", None)
new_images = []
image_new_size=None
#TODO: FIX THE BUG OF NEW SIZE BATCH
# print("DEBUG image_aspect_ratio: ",image_aspect_ratio)
if image_aspect_ratio == 'pad':
for image in images:
image = expand2square(image, tuple(int(x*255) for x in image_processor.image_mean))
image = image_processor.preprocess(image, return_tensors='pt')['pixel_values'][0]
new_images.append(image)
elif image_aspect_ratio == "anyres":
for image in images:
image,image_new_size = process_anyres_image(image, image_processor, model_cfg.image_grid_pinpoints)
new_images.append(image)
elif image_aspect_ratio == "anyres_ui":
# print("DEBUG---: Process As UI")
for image in images:
image,image_new_size = process_anyres_ui_image(image, image_processor,fusion=False)
new_images.append(image)
elif image_aspect_ratio == "anyres_ui_fusion":
for image in images:
# print("DEBUG---: Process As anyres_ui_fusion")
image,image_new_size = process_anyres_ui_image(image, image_processor,fusion=True)
# if image_new_size is not None:
# print("NEW SIZE", image_new_size)
# else:
# print("NEW SIZE IS NONE!!!!")
new_images.append(image)
else:
print(image_aspect_ratio)
raise NotImplementedError
# return image_processor(images, return_tensors='pt')['pixel_values']
# print("LEN new_images",len(new_images))
# if image_new_size is not None:
# print("AFTER: NEW SIZE",image_new_size)
# else:
# print("AFTER: NEW SIZE IS NONE!!!!")
#
# print("TYPE new_images[0]",type(new_images[0]))
# print("len new_images[0]", len(new_images[0]))
# print("new_images[0]", new_images[0])
if all(x.shape == new_images[0].shape for x in new_images):
new_images = torch.stack(new_images, dim=0)
if image_new_size is not None:
# print("RETURN WITH NEW SIZE")
return new_images, image_new_size
else:
# print("RETURN ONLY IMAGE")
return new_images
def tokenizer_image_token(prompt, tokenizer, image_token_index=IMAGE_TOKEN_INDEX, return_tensors=None):
prompt_chunks = [tokenizer(chunk).input_ids for chunk in prompt.split('<image>')]
def insert_separator(X, sep):
return [ele for sublist in zip(X, [sep]*len(X)) for ele in sublist][:-1]
input_ids = []
offset = 0
if len(prompt_chunks) > 0 and len(prompt_chunks[0]) > 0 and prompt_chunks[0][0] == tokenizer.bos_token_id:
offset = 1
input_ids.append(prompt_chunks[0][0])
for x in insert_separator(prompt_chunks, [image_token_index] * (offset + 1)):
input_ids.extend(x[offset:])
if return_tensors is not None:
if return_tensors == 'pt':
return torch.tensor(input_ids, dtype=torch.long)
raise ValueError(f'Unsupported tensor type: {return_tensors}')
return input_ids
def get_model_name_from_path(model_path):
model_path = model_path.strip("/")
model_paths = model_path.split("/")
if model_paths[-1].startswith('checkpoint-'):
return model_paths[-2] + "_" + model_paths[-1]
else:
return model_paths[-1]
class KeywordsStoppingCriteria(StoppingCriteria):
def __init__(self, keywords, tokenizer, input_ids):
self.keywords = keywords
self.keyword_ids = []
self.max_keyword_len = 0
for keyword in keywords:
cur_keyword_ids = tokenizer(keyword).input_ids
if len(cur_keyword_ids) > 1 and cur_keyword_ids[0] == tokenizer.bos_token_id:
cur_keyword_ids = cur_keyword_ids[1:]
if len(cur_keyword_ids) > self.max_keyword_len:
self.max_keyword_len = len(cur_keyword_ids)
self.keyword_ids.append(torch.tensor(cur_keyword_ids))
self.tokenizer = tokenizer
self.start_len = input_ids.shape[1]
def call_for_batch(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
offset = min(output_ids.shape[1] - self.start_len, self.max_keyword_len)
self.keyword_ids = [keyword_id.to(output_ids.device) for keyword_id in self.keyword_ids]
for keyword_id in self.keyword_ids:
truncated_output_ids = output_ids[0, -keyword_id.shape[0]:]
if torch.equal(truncated_output_ids, keyword_id):
return True
outputs = self.tokenizer.batch_decode(output_ids[:, -offset:], skip_special_tokens=True)[0]
for keyword in self.keywords:
if keyword in outputs:
return True
return False
def __call__(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
outputs = []
for i in range(output_ids.shape[0]):
outputs.append(self.call_for_batch(output_ids[i].unsqueeze(0), scores))
return all(outputs)