File size: 4,592 Bytes
08a9c32 |
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 |
import numpy as np
import torch
import torchvision.transforms as T
from PIL import Image
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoModel, AutoTokenizer
from internvl2_patches import InternVLChatModel
import config
# If you want to load a model using multiple GPUs, please refer to the `Multiple GPUs` section.
path = config.path
model = InternVLChatModel.from_pretrained(
path,
torch_dtype=config.dtype,
# low_cpu_mem_usage=True,
use_flash_attn=True,
ignore_mismatched_sizes=True,
revision='7f49802f5bf1e6e3d20b6f69268701c7eb67e037').to(config.device)
tokenizer = AutoTokenizer.from_pretrained('OpenGVLab/InternVL2-4B', trust_remote_code=True, use_fast=False,
revision='7f49802f5bf1e6e3d20b6f69268701c7eb67e037')
tokenizer.padding_side = 'left'
img_context_token_id = tokenizer.convert_tokens_to_ids('<IMG_CONTEXT>')
model.img_context_token_id = img_context_token_id
model.mlp1 = model.mlp1.to(torch.float32)
# model.vision_model.encoder = model.vision_model.encoder.to(torch.float32)
print(model.mlp1,)
params = list(model.mlp1.parameters())# + list(model.vision_model.encoder.parameters())
print(f'Training: {params}')
# we will drop all but last patch & train mlp1; mlp1 will be where we do vector arythmetic and probes.
optimizer = torch.optim.AdamW(params, lr=config.lr)
IMAGENET_MEAN = (0.485, 0.456, 0.406)
IMAGENET_STD = (0.229, 0.224, 0.225)
def build_transform(input_size):
MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
transform = T.Compose([
T.Lambda(lambda img: img.convert('RGB') if img.mode != 'RGB' else img),
T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
T.ToTensor(),
T.Normalize(mean=MEAN, std=STD)
])
return transform
def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
best_ratio_diff = float('inf')
best_ratio = (1, 1)
area = width * height
for ratio in target_ratios:
target_aspect_ratio = ratio[0] / ratio[1]
ratio_diff = abs(aspect_ratio - target_aspect_ratio)
if ratio_diff < best_ratio_diff:
best_ratio_diff = ratio_diff
best_ratio = ratio
elif ratio_diff == best_ratio_diff:
if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
best_ratio = ratio
return best_ratio
def dynamic_preprocess(image, min_num=1, max_num=12, image_size=448, use_thumbnail=False):
orig_width, orig_height = image.size
aspect_ratio = orig_width / orig_height
# calculate the existing image aspect ratio
target_ratios = set(
(i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
i * j <= max_num and i * j >= min_num)
target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
# find the closest aspect ratio to the target
target_aspect_ratio = find_closest_aspect_ratio(
aspect_ratio, target_ratios, orig_width, orig_height, image_size)
# calculate the target width and height
target_width = image_size * target_aspect_ratio[0]
target_height = image_size * target_aspect_ratio[1]
blocks = target_aspect_ratio[0] * target_aspect_ratio[1]
# resize the image
resized_img = image.resize((target_width, target_height))
processed_images = []
for i in range(blocks):
box = (
(i % (target_width // image_size)) * image_size,
(i // (target_width // image_size)) * image_size,
((i % (target_width // image_size)) + 1) * image_size,
((i // (target_width // image_size)) + 1) * image_size
)
# split the image
split_img = resized_img.crop(box)
processed_images.append(split_img)
assert len(processed_images) == blocks
if use_thumbnail and len(processed_images) != 1:
thumbnail_img = image.resize((image_size, image_size))
processed_images.append(thumbnail_img)
return processed_images
# TODO can make a batch process within data pipeline
def load_image(image_file, pil_image=None, input_size=224, max_num=12):
if not pil_image:
pil_image = Image.open(image_file)
image = pil_image.convert('RGB')
transform = build_transform(input_size=input_size)
# images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
pixel_values = [transform(image) for image in [image]]
pixel_values = torch.stack(pixel_values)
return pixel_values
|