Spaces:
Running
on
Zero
Running
on
Zero
File size: 4,224 Bytes
b90b5f6 3461177 b90b5f6 3461177 b90b5f6 3461177 b90b5f6 |
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 |
from PIL import Image
from io import BytesIO
import base64
import torch
from transformers import StoppingCriteria
from llava.constants import IMAGE_TOKEN_INDEX
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 = []
# Handle both single processor and list of processors (multi-scale vision)
if isinstance(image_processor, list):
# Multi-scale: use first processor for preprocessing
processor = image_processor[0]
else:
processor = image_processor
if image_aspect_ratio == 'pad':
for image in images:
image = expand2square(image, tuple(int(x*255) for x in processor.image_mean))
image = processor.preprocess(image, return_tensors='pt')['pixel_values'][0]
new_images.append(image)
else:
return processor(images, return_tensors='pt')['pixel_values']
if all(x.shape == new_images[0].shape for x in new_images):
new_images = torch.stack(new_images, dim=0)
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__(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
assert output_ids.shape[0] == 1, "Only support batch size 1 (yet)" # TODO
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:
if (output_ids[0, -keyword_id.shape[0]:] == keyword_id).all():
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
|