optimum-intel-internal-testing/tiny-random-phi3-vision
021k
1# coding=utf-82# Copyright 2024 Microsoft and the HuggingFace Inc. team. All rights reserved.3#4# Licensed under the Apache License, Version 2.0 (the "License");5# you may not use this file except in compliance with the License.6# You may obtain a copy of the License at7#8# http://www.apache.org/licenses/LICENSE-2.09#10# Unless required by applicable law or agreed to in writing, software11# distributed under the License is distributed on an "AS IS" BASIS,12# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.13# See the License for the specific language governing permissions and14# limitations under the License.15 16"""17Processor class for Phi3-V.18"""19import re20from typing import List, Optional, Union21 22import torch23 24import transformers25from transformers.feature_extraction_utils import BatchFeature26from transformers.image_utils import ImageInput27from transformers.processing_utils import ProcessorMixin28from transformers.tokenization_utils_base import PaddingStrategy, TextInput, TruncationStrategy29from transformers.utils import TensorType30 31 32"""Image processor class for Phi3-V."""33 34from typing import List, Optional, Union35 36import numpy as np37 38from transformers.image_processing_utils import BaseImageProcessor, BatchFeature39from transformers.image_transforms import (40 convert_to_rgb,41)42from transformers.image_utils import (43 OPENAI_CLIP_MEAN,44 OPENAI_CLIP_STD,45 ImageInput,46 make_list_of_images,47 valid_images,48)49from transformers.utils import TensorType, is_vision_available, logging50 51from transformers import AutoImageProcessor52 53logger = logging.get_logger(__name__)54 55 56if is_vision_available():57 from PIL import Image58 59import torch60import torchvision61 62def padding_336(b):63 width, height = b.size64 tar = int(np.ceil(height / 336) * 336)65 top_padding = int((tar - height)/2)66 bottom_padding = tar - height - top_padding67 left_padding = 068 right_padding = 069 b = torchvision.transforms.functional.pad(b, [left_padding, top_padding, right_padding, bottom_padding], fill=[255,255,255])70 71 return b72 73def calc_padded_size(width, height, padding_unit=336): 74 target_height = int(np.ceil(height / padding_unit) * padding_unit) 75 top_padding = int((target_height - height) / 2) 76 bottom_padding = target_height - height - top_padding 77 left_padding = 0 78 right_padding = 0 79 padded_width = width + left_padding + right_padding 80 padded_height = height + top_padding + bottom_padding 81 return padded_width, padded_height 82 83def HD_transform(img, hd_num=16):84 width, height = img.size85 trans = False86 if width < height:87 img = img.transpose(Image.TRANSPOSE)88 trans = True89 width, height = img.size90 ratio = (width/ height)91 scale = 192 while scale*np.ceil(scale/ratio) <= hd_num:93 scale += 194 scale -= 195 new_w = int(scale * 336)96 new_h = int(new_w / ratio)97 98 img = torchvision.transforms.functional.resize(img, [new_h, new_w],)99 img = padding_336(img)100 width, height = img.size101 if trans:102 img = img.transpose(Image.TRANSPOSE)103 104 return img105 106def calc_hd_transform_size(width, height, hd_num=16): 107 transposed = False 108 if width < height: 109 width, height = height, width 110 transposed = True 111 112 ratio = width / height 113 scale = 1 114 while scale * np.ceil(scale / ratio) <= hd_num: 115 scale += 1 116 scale -= 1 117 118 new_width = int(scale * 336) 119 new_height = int(new_width / ratio) 120 121 padded_width, padded_height = calc_padded_size(new_width, new_height) 122 123 if transposed: 124 padded_width, padded_height = padded_height, padded_width 125 126 return padded_width, padded_height 127 128def pad_to_max_num_crops_tensor(images, max_crops=5):129 """130 images: B x 3 x H x W, B<=max_crops131 """132 B, _, H, W = images.shape133 if B < max_crops:134 pad = torch.zeros(max_crops - B, 3, H, W, dtype=images.dtype, device=images.device)135 images = torch.cat([images, pad], dim=0)136 return images137 138 139class Phi3VImageProcessor(BaseImageProcessor):140 r"""141 Constructs a Phi3 image processor. Based on [`CLIPImageProcessor`] with incorporation of additional techniques142 for processing high resolution images as explained in the [InternLM-XComposer2-4KHD](https://arxiv.org/pdf/2404.06512)143 144 Args:145 image_mean (`float` or `List[float]`, *optional*, defaults to `[0.48145466, 0.4578275, 0.40821073]`):146 Mean to use if normalizing the image. This is a float or list of floats the length of the number of147 channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method.148 image_std (`float` or `List[float]`, *optional*, defaults to `[0.26862954, 0.26130258, 0.27577711]`):149 Standard deviation to use if normalizing the image. This is a float or list of floats the length of the150 number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method.151 Can be overridden by the `image_std` parameter in the `preprocess` method.152 do_convert_rgb (`bool`, *optional*, defaults to `True`):153 Whether to convert the image to RGB.154 """155 156 model_input_names = ["pixel_values"]157 158 def __init__(159 self,160 num_crops: int = 1,161 image_mean: Optional[Union[float, List[float]]] = None,162 image_std: Optional[Union[float, List[float]]] = None,163 do_convert_rgb: bool = True,164 **kwargs,165 ) -> None:166 super().__init__(**kwargs)167 self.num_crops = num_crops168 self.image_mean = image_mean if image_mean is not None else OPENAI_CLIP_MEAN169 self.image_std = image_std if image_std is not None else OPENAI_CLIP_STD170 self.do_convert_rgb = do_convert_rgb171 172 def calc_num_image_tokens(173 self, 174 images: ImageInput 175 ):176 """ Calculate the number of image tokens for each image.177 Args:178 images (`ImageInput`):179 Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If180 passing in images with pixel values between 0 and 1, set `do_rescale=False`.181 """182 images = make_list_of_images(images)183 184 if not valid_images(images):185 raise ValueError(186 "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "187 "torch.Tensor, tf.Tensor or jax.ndarray."188 )189 190 images = [image.convert('RGB') for image in images]191 # (H, W, C)192 elems = [HD_transform(im, hd_num = self.num_crops) for im in images] 193 shapes = [[im.size[1], im.size[0]] for im in elems]194 num_img_tokens = [int((h//336*w//336+1)*144 + 1 + (h//336+1)*12) for h, w in shapes]195 return num_img_tokens196 197 def calc_num_image_tokens_from_image_size(self, width, height):198 """199 Calculate the number of image tokens for a given image size.200 Args:201 width (`int`): Width of the image.202 height (`int`): Height of the image.203 """204 new_width, new_height = calc_hd_transform_size(width, height, hd_num=self.num_crops) 205 num_img_tokens = int((new_height // 336 * new_width // 336 + 1) * 144 + 1 + (new_height // 336 + 1) * 12) 206 return num_img_tokens207 208 def preprocess(209 self,210 images: ImageInput,211 image_mean: Optional[Union[float, List[float]]] = None,212 image_std: Optional[Union[float, List[float]]] = None,213 do_convert_rgb: bool = None,214 return_tensors: Optional[Union[str, TensorType]] = None,215 ):216 """217 Args:218 images (`ImageInput`):219 Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If220 passing in images with pixel values between 0 and 1, set `do_rescale=False`.221 image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`):222 Image mean to use for normalization. Only has an effect if `do_normalize` is set to `True`.223 image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`):224 Image standard deviation to use for normalization. Only has an effect if `do_normalize` is set to225 `True`.226 do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`):227 Whether to convert the image to RGB.228 return_tensors (`str` or `TensorType`, *optional*):229 The type of tensors to return. Can be one of:230 - Unset: Return a list of `np.ndarray`.231 - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`.232 - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`.233 - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`.234 - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`.235 """236 image_mean = image_mean if image_mean is not None else self.image_mean237 image_std = image_std if image_std is not None else self.image_std238 do_convert_rgb = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb239 240 images = make_list_of_images(images)241 242 if not valid_images(images):243 raise ValueError(244 "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "245 "torch.Tensor, tf.Tensor or jax.ndarray."246 )247 248 if do_convert_rgb:249 images = [convert_to_rgb(image) for image in images]250 251 image_sizes = []252 img_processor = torchvision.transforms.Compose([253 torchvision.transforms.ToTensor(),254 torchvision.transforms.Normalize(image_mean, image_std)255 ])256 257 # PIL images258 # HD_transform pad images to size of multiiply of 336, 336259 # convert to RGB first260 images = [image.convert('RGB') for image in images]261 elems = [HD_transform(im, hd_num = self.num_crops) for im in images] 262 # tensor transform and normalize263 hd_images = [img_processor(im) for im in elems]264 # create global image 265 global_image = [torch.nn.functional.interpolate(im.unsqueeze(0).float(), size=(336, 336), mode='bicubic',).to(im.dtype) for im in hd_images]266 267 # [(3, h, w)], where h, w is multiple of 336268 shapes = [[im.size(1), im.size(2)] for im in hd_images]269 num_img_tokens = [int(((h//336)*(w//336)+1)*144 + 1 + (h//336+1)*12) for h, w in shapes]270 # reshape to channel dimension -> (num_images, num_crops, 3, 336, 336)271 # (1, 3, h//336, 336, w//336, 336) -> (1, h//336, w//336, 3, 336, 336) -> (h//336*w//336, 3, 336, 336)272 hd_images_reshape = [im.reshape(1, 3, h//336, 336, w//336, 336).permute(0,2,4,1,3,5).reshape(-1, 3, 336, 336).contiguous() for im, (h, w) in zip(hd_images, shapes)]273 # concat global image and local image274 hd_images_reshape = [torch.cat([_global_image] + [_im], dim=0) for _global_image, _im in zip(global_image, hd_images_reshape)]275 276 # pad to max_num_crops277 image_transformed = [pad_to_max_num_crops_tensor(im, self.num_crops+1) for im in hd_images_reshape]278 image_transformed = torch.stack(image_transformed, dim=0)279 image_sizes = [torch.LongTensor(_shapes) for _shapes in shapes]280 padded_images = image_transformed281 image_sizes = shapes282 283 data = {"pixel_values": padded_images, 284 "image_sizes": image_sizes,285 "num_img_tokens": num_img_tokens286 }287 288 return BatchFeature(data=data, tensor_type=return_tensors)289 290AutoImageProcessor.register("Phi3VImageProcessor", Phi3VImageProcessor)291 292transformers.Phi3VImageProcessor = Phi3VImageProcessor 293 294class Phi3VProcessor(ProcessorMixin):295 r"""296 Constructs a Phi3-V processor which wraps a Phi3-V image processor and a LLaMa tokenizer into a single processor.297 298 [`Phi3VProcessor`] offers all the functionalities of [`Phi3VImageProcessor`] and [`LlamaTokenizerFast`]. See the299 [`~Phi3VProcessor.__call__`] and [`~Phi3VProcessor.decode`] for more information.300 301 Args:302 image_processor ([`Phi3VImageProcessor`], *optional*):303 The image processor is a required input.304 tokenizer ([`LlamaTokenizerFast`], *optional*):305 The tokenizer is a required input.306 """307 308 attributes = ["image_processor", "tokenizer"]309 image_processor_class = "Phi3VImageProcessor"310 tokenizer_class = ("LlamaTokenizer", "LlamaTokenizerFast")311 special_image_token = "<|image|>"312 valid_kwargs = ["chat_template"]313 314 def __init__(self, image_processor, tokenizer, **kwargs):315 self.image_processor = image_processor316 self.tokenizer = tokenizer317 self.num_img_tokens = image_processor.num_img_tokens318 self.img_tokens = [f"<|image_{i+1}|>" for i in range(1000000)]319 320 def __call__(321 self,322 text: Union[TextInput, List[TextInput]],323 images: ImageInput = None,324 padding: Union[bool, str, PaddingStrategy] = False,325 truncation: Union[bool, str, TruncationStrategy] = None,326 max_length=None,327 return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,328 ) -> BatchFeature:329 """330 Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`331 and `kwargs` arguments to LlamaTokenizerFast's [`~LlamaTokenizerFast.__call__`] if `text` is not `None` to encode332 the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to333 Phi3ImageProcessor's [`~Phi3ImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring334 of the above two methods for more information.335 336 Args:337 text (`str`, `List[str]`, `List[List[str]]`):338 The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings339 (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set340 `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).341 images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):342 The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch343 tensor. Both channels-first and channels-last formats are supported.344 padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):345 Select a strategy to pad the returned sequences (according to the model's padding side and padding346 index) among:347 - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single348 sequence if provided).349 - `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum350 acceptable input length for the model if that argument is not provided.351 - `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different352 lengths).353 max_length (`int`, *optional*):354 Maximum length of the returned list and optionally padding length (see above).355 truncation (`bool`, *optional*):356 Activates truncation to cut input sequences longer than `max_length` to `max_length`.357 return_tensors (`str` or [`~utils.TensorType`], *optional*):358 If set, will return tensors of a particular framework. Acceptable values are:359 360 - `'tf'`: Return TensorFlow `tf.constant` objects.361 - `'pt'`: Return PyTorch `torch.Tensor` objects.362 - `'np'`: Return NumPy `np.ndarray` objects.363 - `'jax'`: Return JAX `jnp.ndarray` objects.364 365 Returns:366 [`BatchFeature`]: A [`BatchFeature`] with the following fields:367 368 - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.369 - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when370 `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not371 `None`).372 - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.373 """374 if images is not None:375 image_inputs = self.image_processor(images, return_tensors=return_tensors)376 else:377 image_inputs = {}378 inputs = self._convert_images_texts_to_inputs(image_inputs, text, padding=padding, truncation=truncation, max_length=max_length, return_tensors=return_tensors)379 return inputs380 381 def calc_num_image_tokens(self, images: ImageInput):382 """ Calculate the number of image tokens for each image.383 Args:384 images (`ImageInput`):385 Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If386 passing in images with pixel values between 0 and 1, set `do_rescale=False`.387 """388 return self.image_processor.calc_num_image_tokens(images)389 390 def calc_num_image_tokens_from_image_size(self, width, height):391 """ Calculate the number of image token for an image with given width and height.392 Args:393 width (`int`):394 Width of the image.395 height (`int`):396 Height of the image.397 """398 return self.image_processor.calc_num_image_tokens_from_image_size(width, height)399 400 401 @property 402 def special_image_token_id(self):403 return self.tokenizer.convert_tokens_to_ids(self.special_image_token)404 405 def get_special_image_token_id(self):406 return self.tokenizer.convert_tokens_to_ids(self.special_image_token)407 408 def _convert_images_texts_to_inputs(self, images, texts, padding=False, truncation=None, max_length=None, return_tensors=None):409 410 if not len(images):411 model_inputs = self.tokenizer(texts, return_tensors=return_tensors, padding=padding, truncation=truncation, max_length=max_length)412 return BatchFeature(data={**model_inputs})413 414 pattern = r"<\|image_\d+\|>"415 prompt_chunks = [self.tokenizer(chunk).input_ids for chunk in re.split(pattern, texts)] 416 417 if 'num_img_tokens' in images:418 num_img_tokens = images['num_img_tokens']419 else:420 assert 'num_crops' in images, 'num_crops must be provided in images if num_img_tokens is not provided'421 num_crops = images['num_crops']422 num_img_tokens = [_num_crops * self.num_img_tokens for _num_crops in num_crops] 423 424 images, image_sizes = images['pixel_values'], images['image_sizes']425 426 # image_tags needs to start from 1 to n427 image_tags = re.findall(pattern, texts) 428 # image_ids = [int(s.split("|")[1].split("_")[-1]) * -1 for s in image_tags]429 # image_ids_pad = [[iid]*num_img_tokens[i] for i, iid in enumerate(image_ids)]430 image_ids = [int(s.split("|")[1].split("_")[-1]) for s in image_tags]431 unique_image_ids = sorted(list(set(image_ids)))432 # image_ids must start from 1, and must be continuous int, e.g. [1, 2, 3], cannot be [1, 4, 5]433 # check the condition434 assert unique_image_ids == list(range(1, len(unique_image_ids)+1)), f"image_ids must start from 1, and must be continuous int, e.g. [1, 2, 3], cannot be {unique_image_ids}"435 # total images must be the same as the number of image tags436 assert len(unique_image_ids) == len(images), f"total images must be the same as the number of image tags, got {len(unique_image_ids)} image tags and {len(images)} images"437 438 image_ids_pad = [[-iid]*num_img_tokens[iid-1] for iid in image_ids]439 440 def insert_separator(X, sep_list):441 if len(X) > len(sep_list):442 sep_list.append([])443 return [ele for sublist in zip(X, sep_list) for ele in sublist]444 input_ids = []445 offset = 0 446 for x in insert_separator(prompt_chunks, image_ids_pad):447 input_ids.extend(x[offset:])448 449 input_ids = torch.tensor(input_ids, dtype=torch.long).unsqueeze(0)450 attention_mask = (input_ids > -1000000).to(torch.long)451 452 return BatchFeature(data={"input_ids": input_ids,453 "attention_mask": attention_mask,454 "pixel_values": images, 455 "image_sizes": image_sizes})456 457 458 # Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama459 def batch_decode(self, *args, **kwargs):460 """461 This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please462 refer to the docstring of this method for more information.463 """464 return self.tokenizer.batch_decode(*args, **kwargs)465 466 # Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama467 def decode(self, *args, **kwargs):468 """469 This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to470 the docstring of this method for more information.471 """472 return self.tokenizer.decode(*args, **kwargs)473 474 @property475 # Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names476 def model_input_names(self):477 tokenizer_input_names = self.tokenizer.model_input_names478 image_processor_input_names = self.image_processor.model_input_names479 return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))