import ast import io import json import logging import math import os import random import sys import braceexpand from dataclasses import dataclass from multiprocessing import Value import numpy.lib.format import numpy as np import pandas as pd import torch import torchvision.datasets as datasets import webdataset as wds from PIL import Image from torch.utils.data import Dataset, DataLoader, SubsetRandomSampler, IterableDataset, get_worker_info from torch.utils.data.distributed import DistributedSampler from torchvision.transforms import ToTensor from tqdm import tqdm from webdataset.filters import _shuffle from webdataset.tariterators import base_plus_ext, url_opener, tar_file_expander, valid_sample from open_clip import get_tokenizer from open_clip.factory import HF_HUB_PREFIX from training.params import parse_args from data.process_text import load_and_transform_text from data.process_video import get_video_transform from data.process_audio import get_audio_transform from data.process_depth import get_depth_transform from data.process_thermal import get_thermal_transform import pdb try: import horovod.torch as hvd except ImportError: hvd = None class SharedEpoch: def __init__(self, epoch: int = 0): self.shared_epoch = Value('i', epoch) def set_value(self, epoch): self.shared_epoch.value = epoch def get_value(self): return self.shared_epoch.value @dataclass class DataInfo: dataloader: DataLoader sampler: DistributedSampler = None shared_epoch: SharedEpoch = None def set_epoch(self, epoch): if self.shared_epoch is not None: self.shared_epoch.set_value(epoch) if self.sampler is not None and isinstance(self.sampler, DistributedSampler): self.sampler.set_epoch(epoch) def expand_urls(urls, weights=None): if weights is None: expanded_urls = wds.shardlists.expand_urls(urls) return expanded_urls, None if isinstance(urls, str): urllist = urls.split("::") weights = weights.split('::') assert len(weights) == len(urllist), \ f"Expected the number of data components ({len(urllist)}) and weights({len(weights)}) to match." weights = [float(weight) for weight in weights] all_urls, all_weights = [], [] for url, weight in zip(urllist, weights): expanded_url = list(braceexpand.braceexpand(url)) expanded_weights = [weight for _ in expanded_url] all_urls.extend(expanded_url) all_weights.extend(expanded_weights) return all_urls, all_weights else: all_urls = list(urls) return all_urls, weights def get_dataset_size(shards): shards_list, _ = expand_urls(shards) dir_path = os.path.dirname(shards_list[0]) sizes_filename = os.path.join(dir_path, 'sizes.json') len_filename = os.path.join(dir_path, '__len__') if os.path.exists(sizes_filename): sizes = json.load(open(sizes_filename, 'r')) total_size = sum([int(sizes[os.path.basename(shard)]) for shard in shards_list]) elif os.path.exists(len_filename): # FIXME this used to be eval(open(...)) but that seemed rather unsafe total_size = ast.literal_eval(open(len_filename, 'r').read()) else: total_size = None # num samples undefined # some common dataset sizes (at time of authors last download) # CC3M (train): 2905954 # CC12M: 10968539 # LAION-400M: 407332084 # LAION-2B (english): 2170337258 num_shards = len(shards_list) return total_size, num_shards def count_samples(dataloader): os.environ["WDS_EPOCH"] = "0" n_elements, n_batches = 0, 0 for images, texts in dataloader: n_batches += 1 n_elements += len(images) assert len(images) == len(texts) return n_elements, n_batches def filter_no_caption_or_no_image(sample): has_caption = ('raw.txt' in sample and 'mplug.txt' in sample and 'polish_mplug.txt' in sample and 'ofa3.txt' in sample) has_image = ('frm7.jpg' in sample and 'tml0.jpg' in sample and 'dep0.npy' in sample) return has_caption and has_image def log_and_continue(exn): """Call in an exception handler to ignore any exception, issue a warning, and continue.""" logging.warning(f'Handling webdataset error ({repr(exn)}). Ignoring.') return True def group_by_keys_nothrow(data, keys=base_plus_ext, lcase=True, suffixes=None, handler=None): """Return function over iterator that groups key, value pairs into samples. :param keys: function that splits the key into key and extension (base_plus_ext) :param lcase: convert suffixes to lower case (Default value = True) """ current_sample = None for filesample in data: assert isinstance(filesample, dict) fname, value = filesample["fname"], filesample["data"] prefix, suffix = keys(fname) if prefix is None: continue if lcase: suffix = suffix.lower() # FIXME webdataset version throws if suffix in current_sample, but we have a potential for # this happening in the current LAION400m dataset if a tar ends with same prefix as the next # begins, rare, but can happen since prefix aren't unique across tar files in that dataset if current_sample is None or prefix != current_sample["__key__"] or suffix in current_sample: if valid_sample(current_sample): yield current_sample current_sample = dict(__key__=prefix, __url__=filesample["__url__"]) if suffixes is None or suffix in suffixes: current_sample[suffix] = value if valid_sample(current_sample): yield current_sample def tarfile_to_samples_nothrow(src, handler=log_and_continue): # NOTE this is a re-impl of the webdataset impl with group_by_keys that doesn't throw streams = url_opener(src, handler=handler) files = tar_file_expander(streams, handler=handler) samples = group_by_keys_nothrow(files, handler=handler) return samples def pytorch_worker_seed(increment=0): """get dataloader worker seed from pytorch""" worker_info = get_worker_info() if worker_info is not None: # favour using the seed already created for pytorch dataloader workers if it exists seed = worker_info.seed if increment: # space out seed increments so they can't overlap across workers in different iterations seed += increment * max(1, worker_info.num_workers) return seed # fallback to wds rank based seed return wds.utils.pytorch_worker_seed() _SHARD_SHUFFLE_SIZE = 200 _SHARD_SHUFFLE_INITIAL = 50 _SAMPLE_SHUFFLE_SIZE = 500 _SAMPLE_SHUFFLE_INITIAL = 100 class detshuffle2(wds.PipelineStage): def __init__( self, bufsize=1000, initial=100, seed=0, epoch=-1, ): self.bufsize = bufsize self.initial = initial self.seed = seed self.epoch = epoch def run(self, src): if isinstance(self.epoch, SharedEpoch): epoch = self.epoch.get_value() else: # NOTE: this is epoch tracking is problematic in a multiprocess (dataloader workers or train) # situation as different workers may wrap at different times (or not at all). self.epoch += 1 epoch = self.epoch rng = random.Random() if self.seed < 0: # If seed is negative, we use the worker's seed, this will be different across all nodes/workers seed = pytorch_worker_seed(epoch) else: # This seed to be deterministic AND the same across all nodes/workers in each epoch seed = self.seed + epoch rng.seed(seed) return _shuffle(src, self.bufsize, self.initial, rng) class ResampledShards2(IterableDataset): """An iterable dataset yielding a list of urls.""" def __init__( self, urls, weights=None, nshards=sys.maxsize, worker_seed=None, deterministic=False, epoch=-1, ): """Sample shards from the shard list with replacement. :param urls: a list of URLs as a Python list or brace notation string """ super().__init__() urls, weights = expand_urls(urls, weights) self.urls = urls self.weights = weights if self.weights is not None: assert len(self.urls) == len(self.weights), \ f"Number of urls {len(self.urls)} and weights {len(self.weights)} should match." assert isinstance(self.urls[0], str) self.nshards = nshards self.rng = random.Random() self.worker_seed = worker_seed self.deterministic = deterministic self.epoch = epoch def __iter__(self): """Return an iterator over the shards.""" if isinstance(self.epoch, SharedEpoch): epoch = self.epoch.get_value() else: # NOTE: this is epoch tracking is problematic in a multiprocess (dataloader workers or train) # situation as different workers may wrap at different times (or not at all). self.epoch += 1 epoch = self.epoch if self.deterministic: # reset seed w/ epoch if deterministic if self.worker_seed is None: # pytorch worker seed should be deterministic due to being init by arg.seed + rank + worker id seed = pytorch_worker_seed(epoch) else: seed = self.worker_seed() + epoch self.rng.seed(seed) for _ in range(self.nshards): if self.weights is None: yield dict(url=self.rng.choice(self.urls)) else: yield dict(url=self.rng.choices(self.urls, weights=self.weights, k=1)[0]) class Decode: def __init__(self, args=None): self.num_frames = args.num_frames self.text_type = args.text_type self.chatgpt = self.text_type == 'polish_mplug' self.title = self.text_type == 'raw' self.clip_type = args.clip_type self.tokenizer = get_tokenizer(HF_HUB_PREFIX + args.model, cache_dir=args.cache_dir) self.video_transform = get_video_transform(args) self.audio_transform = get_audio_transform(args) self.depth_transform = get_depth_transform(args) self.thermal_transform = get_thermal_transform(args) def __call__(self, sample): input_ids, attention_mask = self.get_text(sample[f"{self.text_type}.txt"], chatgpt=self.chatgpt, title=self.title) if self.clip_type == 'vl': matched_modality = self.get_video([sample[f"frm{i}.jpg"] for i in range(self.num_frames)]) elif self.clip_type == 'al': matched_modality = self.get_audio() elif self.clip_type == 'dl': matched_modality = self.get_depth(sample[f"dep0.npy"]) elif self.clip_type == 'tl': matched_modality = self.get_thermal(sample[f"tml0.jpg"]) # matched_modality = self.get_thermal(sample[f"tml{random.randint(0, 7)}.jpg"]) else: raise ValueError return matched_modality, input_ids, attention_mask def get_video(self, frames): video_data = [] for frame in frames: with io.BytesIO(frame) as stream: img = Image.open(stream) img.load() assert min(img.size) == 256 result = ToTensor()(img) video_data.append(result) video_data = torch.stack(video_data, dim=1) # video_data torch.Size([3, 8, 455, 256]) # video_outputs torch.Size([3, 8, 224, 224]) video_outputs = self.video_transform(video_data) return video_outputs def get_text(self, text, chatgpt=True, title=False): text = text.decode("utf-8") if chatgpt: assert text.startswith('In the video, ') text = text[14:] tokens = load_and_transform_text(text, self.tokenizer, title=title) return tokens['input_ids'], tokens['attention_mask'] def get_audio(self): raise NotImplementedError def get_depth(self, depth): stream = io.BytesIO(depth) img = numpy.lib.format.read_array(stream) depth = self.depth_transform(img) return depth def get_thermal(self, thermal): with io.BytesIO(thermal) as stream: img = Image.open(stream) img.load() thermal = self.thermal_transform(img) return thermal def get_wds_dataset(args, is_train, epoch=0, floor=False): input_shards = args.train_data if is_train else args.val_data assert input_shards is not None resampled = getattr(args, 'dataset_resampled', False) and is_train num_shards = None if is_train: if args.train_num_samples is not None: num_samples = args.train_num_samples else: num_samples, num_shards = get_dataset_size(input_shards) if not num_samples: raise RuntimeError( 'Currently, the number of dataset samples must be specified for the training dataset. ' 'Please specify it via `--train-num-samples` if no dataset length info is present.') else: # Eval will just exhaust the iterator if the size is not specified. num_samples = args.val_num_samples or 0 shared_epoch = SharedEpoch(epoch=epoch) # create a shared epoch store to sync epoch to dataloader worker proc if resampled: pipeline = [ResampledShards2( input_shards, weights=args.train_data_upsampling_factors, deterministic=True, epoch=shared_epoch, )] else: assert args.train_data_upsampling_factors is None, \ "--train_data_upsampling_factors is only supported when sampling with replacement (with --dataset-resampled)." pipeline = [wds.SimpleShardList(input_shards)] # at this point we have an iterator over all the shards if is_train: if not resampled: pipeline.extend([ detshuffle2( bufsize=_SHARD_SHUFFLE_SIZE, initial=_SHARD_SHUFFLE_INITIAL, seed=args.seed, epoch=shared_epoch, ), wds.split_by_node, wds.split_by_worker, ]) pipeline.extend([ # at this point, we have an iterator over the shards assigned to each worker at each node tarfile_to_samples_nothrow, # wds.tarfile_to_samples(handler=log_and_continue), wds.shuffle( bufsize=_SAMPLE_SHUFFLE_SIZE, initial=_SAMPLE_SHUFFLE_INITIAL, ), ]) else: pipeline.extend([ wds.split_by_worker, # at this point, we have an iterator over the shards assigned to each worker wds.tarfile_to_samples(handler=log_and_continue), ]) pipeline.extend([ wds.select(filter_no_caption_or_no_image), # wds.decode("pilrgb", handler=log_and_continue), # wds.rename(image="jpg;png;jpeg;webp", text="txt"), # wds.map_dict(image=preprocess_img, text=lambda text: tokenizer(text)[0]), # wds.to_tuple("image", "text"), wds.map(Decode(args), handler=log_and_continue), wds.batched(args.batch_size, partial=not is_train) ]) dataset = wds.DataPipeline(*pipeline) if is_train: if not resampled: num_shards = num_shards or len(expand_urls(input_shards)[0]) assert num_shards >= args.workers * args.world_size, 'number of shards must be >= total workers' # roll over and repeat a few samples to get same number of full batches on each node round_fn = math.floor if floor else math.ceil global_batch_size = args.batch_size * args.world_size num_batches = round_fn(num_samples / global_batch_size) num_workers = max(1, args.workers) num_worker_batches = round_fn(num_batches / num_workers) # per dataloader worker num_batches = num_worker_batches * num_workers num_samples = num_batches * global_batch_size dataset = dataset.with_epoch(num_worker_batches) # each worker is iterating over this else: # last batches are partial, eval is done on single (master) node num_batches = math.ceil(num_samples / args.batch_size) dataloader = wds.WebLoader( dataset, batch_size=None, shuffle=False, num_workers=args.workers, persistent_workers=args.workers > 0, ) # FIXME not clear which approach is better, with_epoch before vs after dataloader? # hoping to resolve via https://github.com/webdataset/webdataset/issues/169 # if is_train: # # roll over and repeat a few samples to get same number of full batches on each node # global_batch_size = args.batch_size * args.world_size # num_batches = math.ceil(num_samples / global_batch_size) # num_workers = max(1, args.workers) # num_batches = math.ceil(num_batches / num_workers) * num_workers # num_samples = num_batches * global_batch_size # dataloader = dataloader.with_epoch(num_batches) # else: # # last batches are partial, eval is done on single (master) node # num_batches = math.ceil(num_samples / args.batch_size) # add meta-data to dataloader instance for convenience dataloader.num_batches = num_batches dataloader.num_samples = num_samples return DataInfo(dataloader=dataloader, shared_epoch=shared_epoch) def get_data(args, epoch=0): data = {} data["train"] = get_wds_dataset(args, is_train=True, epoch=epoch) return data if __name__ == '__main__': args = parse_args(sys.argv[1:]) args.workers = 10 args.batch_size = 16 args.world_size = 1 args.num_frames = 8 args.clip_type = 'vl' args.model = "laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K" args.train_data = '/apdcephfs_cq3/share_1311970/lb/vat2webdata/check_8frm_title_ofa_polishmplug_1tml_1dep/{00000..03020}.tar' args.train_num_samples = 10_000 args.dataset_type = 'webdataset' data = get_data(args, epoch=0) data['train'].set_epoch(0) # set epoch in process safe manner via sampler or shared_epoch dataloader = data['train'].dataloader num_batches_per_epoch = dataloader.num_batches // args.accum_freq print(num_batches_per_epoch) for i, batch in enumerate(tqdm(dataloader)): images, input_ids, attention_mask = batch # print(images.shape, input_ids.shape, attention_mask.shape) # break