import os import gc import torch output_filtering = lambda x, model: x.split(model.prompt_rule["test_start"])[-1].split(model.prompt_rule["test_end"])[0].strip() def memory_optimization(): # memory deallocation gc.collect() # removing cache torch.cuda.empty_cache() def str2bool(v): if v.lower() in ('yes', 'true', 't', 'y', '1'): return True elif v.lower() in ('no', 'false', 'f', 'n', '0'): return False else: assert False def freeze_model(model): for param in model.parameters(): param.requires_grad=False def switching_model(model, updating_param): if updating_param == 'all': for name, param in model.named_parameters(): param.requires_grad=True return for name, param in model.named_parameters(): if 'float' in str(param.dtype): if sum([up_param in name for up_param in updating_param]): param.requires_grad=True else: param.requires_grad=False def weight_upload(tensor_dict, model): used_name = [] for name, param in tensor_dict.items(): split_name = name.split('.') traversal = model for module_name in split_name: traversal = getattr(traversal, module_name) # logging # print(f'{name}: {(traversal==param.to(traversal.device)).sum()}/{(traversal!=param.to(traversal.device)).sum()}') setattr(traversal, 'data', param.to(traversal.device)) used_name.append(name) for name in used_name: del tensor_dict[name] def find_special_token(string, special_token): start = 0 while True: start = string.find(special_token, start) if start == -1: return yield start start += len(special_token) # use start += 1 to find overlapping matches def add_bundle_tokens(input_string, special_token, num): # number of special tokens in input_string num_special_tokens = len(list(find_special_token(input_string, special_token))) # No special token -> return the raw if not num_special_tokens: return input_string result = "" index = 0 while index < len(input_string): if input_string[index:index + len(special_token)] == special_token: result += special_token * num index += len(special_token) else: result += input_string[index] index += 1 assert len(list(find_special_token(result, special_token))) == num_special_tokens * num return result def make_instruction(question, dataset, prompt_rule): system_prompt = make_human_string("You are AI model created by Byung-Kwan Lee, Ph.D. candidate, KAIST EE, of which AI model name is TroL (Traversal of Layers).", "You must give helpful, detailed, and polite answers to the user's questions", split=' ') if dataset != "mmmu" and dataset != "mathverse" and dataset != "hallusionbench" and dataset != "demo": question = "" + question if dataset in ["sqa", "mmbench", "mmbench_cn", "mmbench_dev", "mmbench_cn_dev", "seed", "qbench", "ai2d", "mmstar"]: question = question + "\nAnswer with the option's letter from the given choices directly." elif dataset in ["vqav2", "gqa", "pope", "chartqa"]: question = question + "\nAnswer the question using a single word or phrase." elif dataset in ["vizwiz"]: question = question + "\nWhen the provided information is insufficient, respond with 'Unanswerable'. Answer the question using a single word or phrase." elif dataset in ["mmmu"]: if "A." in question: question = question + "\nAnswer with the option's letter from the given choices directly." else: question = question + "\nAnswer the question using a single word or phrase." elif dataset in ["hallusionbench"]: if "Please answer yes or no." not in question: question = question + "\nPlease answer yes or no." qa_prompt = make_human_string(prompt_rule["system_start"]+system_prompt+prompt_rule["system_end"], prompt_rule["user_start"]+question+prompt_rule["user_end"], prompt_rule["assistant_start"], split=prompt_rule["split"]) return qa_prompt def make_human_string(*args, split): out = '' for i, arg in enumerate(args): out += arg if i != len(args)-1: out += split return out def get_max_new_tokens(data_name): if data_name.lower() in ["mme", "pope", "sqa", "mmbench", "mmbench_cn", "mmbench_dev","mmbench_cn_dev", "seed", "qbench", "ai2d", "mmstar", "vqav2", "gqa", "chartqa", "hallusionbench", "textvqa", "mmmu"]: return 5 if data_name.lower() in ["llava", "mm-vet"]: return 1024 else: return 512 def pixel_shuffle(x, scale_factor=0.5): n, w, h, c = x.size() # N, W, H, C --> N, W, H * scale, C // scale x = x.view(n, w, int(h * scale_factor), int(c / scale_factor)) # N, W, H * scale, C // scale --> N, H * scale, W, C // scale x = x.permute(0, 2, 1, 3).contiguous() # N, H * scale, W, C // scale --> N, H * scale, W * scale, C // (scale ** 2) x = x.view(n, int(h * scale_factor), int(w * scale_factor), int(c / (scale_factor * scale_factor))) x = x.permute(0, 2, 1, 3).contiguous() return x import torchvision.transforms as T from torchvision.transforms.functional import InterpolationMode 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 dynamic_transform = build_transform(input_size=448) 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=6, image_size=448, use_thumbnail=True): from torchvision.transforms.functional import to_pil_image image = to_pil_image(image) 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