evaluation/internvl.py

import math
import torch
import torchvision.transforms as T
from PIL import Image
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoConfig, AutoTokenizer, AutoModel
import os
from utils.utils import COUNTING_BASE_PROMPT, RELATION_BASE_PROMPT, parse_model_answer, eval_loop, eval_pipeline

current_dir = os.path.dirname(os.path.abspath(__file__))
parent_dir = os.path.split(current_dir)[0]
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

    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])

    target_aspect_ratio = find_closest_aspect_ratio(
        aspect_ratio, target_ratios, orig_width, orig_height, image_size)

    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]

    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_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


def load_image(image_file, input_size=448, max_num=12):
    image = Image.open(image_file).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 images]
    pixel_values = torch.stack(pixel_values)
    return pixel_values


def split_model(path):
    device_map = {}
    world_size = torch.cuda.device_count()
    config = AutoConfig.from_pretrained(path, trust_remote_code=True)
    num_layers = config.llm_config.num_hidden_layers
    num_layers_per_gpu = math.ceil(num_layers / (world_size - 0.5))
    num_layers_per_gpu = [num_layers_per_gpu] * world_size
    num_layers_per_gpu[0] = math.ceil(num_layers_per_gpu[0] * 0.5)
    layer_cnt = 0
    for i, num_layer in enumerate(num_layers_per_gpu):
        for j in range(num_layer):
            device_map[f'language_model.model.layers.{layer_cnt}'] = i
            layer_cnt += 1
    device_map['vision_model'] = 0
    device_map['mlp1'] = 0
    device_map['language_model.model.tok_embeddings'] = 0
    device_map['language_model.model.embed_tokens'] = 0
    device_map['language_model.output'] = 0
    device_map['language_model.model.norm'] = 0
    device_map['language_model.model.rotary_emb'] = 0
    device_map['language_model.lm_head'] = 0
    device_map[f'language_model.model.layers.{num_layers - 1}'] = 0
    return device_map


def process(model,  item, task_type="counting", **kwargs):
    """
    :param model:
    :param item:
    :param task_type:
    :param kwargs:
    :return:
    """
    generation_config = kwargs.get('generation_config')
    tokenizer = kwargs.get('tokenizer')
    question = item['question']
    image_path = os.path.join(parent_dir, item['image_path'])
    if task_type == "counting":
        formatted_question = COUNTING_BASE_PROMPT + question
    else:  # relation task
        formatted_question = RELATION_BASE_PROMPT + question
    pixel_values = load_image(image_path, max_num=12).to(torch.bfloat16).cuda()
    model_answer = model.chat(tokenizer, pixel_values, formatted_question, generation_config )
    parsed_answer = parse_model_answer(model_answer, task_type)
    result = {'model_answer': model_answer, 'parsed_answer': parsed_answer}
    return result


def main():
    """
    InternVL3
    """
    model_name_or_path = 'OpenGVLab/InternVL3-2B'
    model_name = model_name_or_path.split('/')[-1]
    cache_dir = './cache/'
    device_map = split_model(model_name_or_path)
    model = AutoModel.from_pretrained(
        model_name_or_path,
        torch_dtype=torch.bfloat16,
        load_in_8bit=False,
        low_cpu_mem_usage=True,
        use_flash_attn=True,
        trust_remote_code=True,
        cache_dir=cache_dir,
        device_map=device_map).eval()

    tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True, use_fast=False)
    generation_config = dict(max_new_tokens=1024, do_sample=True, pad_token_id=tokenizer.convert_tokens_to_ids(tokenizer.pad_token))
    params = {
        'model': model,
        'tokenizer': tokenizer,
        'generation_config': generation_config,
        'process_fn': process,
    }
    eval_pipeline(model_name, current_dir, params)


if __name__ == "__main__":
    main()