Run Effective Large Batch Contrastive Learning on Limited Memory GPU

Overview

Gradient Cache

Gradient Cache is a simple technique for unlimitedly scaling contrastive learning batch far beyond GPU memory constraint. This means training that used to take heavy hardware, e.g. 8 V100 GPU, can be done on a single GPU. In addition, Gradient Cache allow users to replace big RAM GPU with much more cost efficient high FLOP low RAM cards.

This repo holds a generic Pytorch implementation of Gradient Cache described in our paper Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup .

@inproceedings{gao2021scaling,
     title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup},
     author={Luyu Gao, Yunyi Zhang, Jiawei Han, Jamie Callan},
     booktitle ={Proceedings of the 6th Workshop on Representation Learning for NLP},
     year={2021},
}

Gradient Cache has also been integrated into dense passage retrieval (DPR). Checkout our GC-DPR toolkit.

Installation

The package depends only on pytorch>=1.6. To install, clone this repo and run pip.

git clone https://github.com/luyug/GradCache
cd GradCache
pip install .

For development,

pip install --editable .

Usage

Gradient caching functionalities are implemented in GradCache class. If you are developing a new project instead of patching an old one, also checkout our functional approach for a effort reduced approach.

Initialization

The class's __init__ method defines the cache and has several functional parameters *_fn for easy adjust of model behaviors. Alternatively you can also sub-class GradCache.

grad_cache.GradCache(  
  models: List[nn.Module],  
  chunk_sizes: Union[int, List[int]],  
  loss_fn: Callable[..., Tensor],  
  split_input_fn: Callable[[Any, int], Any] = None,  
  get_rep_fn: Callable[..., Tensor] = None,  
  fp16: bool = False,  
  scaler: GradScaler = None,  
)

models - A list of encoder models to be updated with with the Gradient Cache.

chunk_sizes - An integer indicating chunk size. Or a list of integers of chunk size for each model. This controls for each model the sub-batch size to run forward-backward pass and should be set based on available GPU memory. A value too small will leave the GPU under utilized.

loss_fn - A loss function that takes representation tensors of number equal to number of models in models and arbitrary numbers of keyword arguments. It should compute loss based on the input tensors, and in no case modify the input tensors' relations in the autograd graph, which are later relied upon to create the gradient cache.

split_input_fn - An optional function that split generic model input into chunks based on defined chunk_sizes. If not provided, this class will try its best to split the inputs of supported types. See split_inputs function.

get_rep_fn - An optional function that takes generic model output and return representation tensors. If not provided, the generic output is assumed to be the representation tensor.

fp16 - If True, run mixed precision training, which requires scaler to also be set.

scaler - A GradScaler object for automatic mixed precision training.

Cache Gradient Step

To run a cached gradient computatoin step, call cache_step function,

cache_step(  
  *model_inputs,  
  no_sync_except_last: bool = False,  
  **loss_kwargs  
)

Run a single gradient cache step. Upon function return, updates are computed for each model in self.models with gradient populated on the weights, as if the model_inputs are run as a huge single batch on sufficiently large hardware. Calling an GradCache object with __call__ will also invoke this function.

model_inputs - List of inputs to each encoder model. Should be in similar order as self.models.

no_sync_except_last - If True, under distributed setup, for each model, only trigger gradient reduction across processes for the last sub-batch's forward-backward pass. This could come in handy when dealing with a) large model, and/or b) non trivial number of sub-batches.

loss_kwargs - Additional keyword arguments to the loss function loss_fn. This is intended to enable flexible loss computation (thanks to dynamic graph in Pytorch) such as reduction, weighting, etc. Potentially, using loss_kwargs you can incorporate outputs from those encoder models not tracked by the cache.

Return - loss, the current steps loss scaler tensor (detached from the graph).

Natively Supported Input Types

  • x: Tensor - will be passed in as model(x)
  • x: List[Tensor] - will be passed in as model(*x)
  • x: Dict[str, Tensor] (or UserDict[str, Tensor]) - will be passed in as model(**x)
  • x: Tuple[List[Tensor], Dict[str, Tensor]] - will be passed in as model(*x[0], **x[1])

Other generic input are not fully supported, we perform model call using the following heuristics,

  • x: List[Any] - will be passed in as model(*x)
  • x: Dict[str, Any] - will be passed in as model(**x)
  • x: Tuple[List[Any], Dict[str, Any]] - will be passed in as model(*x[0], **x[1])

To run with them, split_input_fn should be specified during cache initialization to break these inputs into smaller batches. In some rare cases, you may also need to override get_input_tensors when its heuristic can not grab enough tensors that covers all cuda devices that hold some tensors in the input.

Example Usage with Huggingface Transformers

Learning a Bi-encoder

Say we want to learn a embedding space of labels and text. Consider the following four pairs. (In practice, you will have many more and much longer text entries.)

labels = ['fruit', 'meat', 'school', 'company']
texts = [
  'this is an apple', 
  'steak should be cooked medium rare', 
  'cmu is pittsburgh', 
  'apple sells laptop'
]

Initialize our encoder models,

from transformers import AutoTokenizer, TFAutoModel
tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
encoder1 = AutoModel.from_pretrained("bert-base-uncased").cuda()
encoder2 = AutoModel.from_pretrained("bert-base-uncased").cuda()

Initialize the GradCache object,

from grad_cache import GradCache
from grad_cache.loss import SimpleContrastiveLoss

loss_fn = SimpleContrastiveLoss()
gc = GradCache(
  models=[encoder1, encoder2], 
  chunk_sizes=2, 
  loss_fn=loss_fn, 
  get_rep_fn=lambda v: v.pooler_output
)

Here we use the get_rep_fn argument to specify a function that takes generic Huggingface model output and return the actual representation tensor.

Create model input,

xx = tokenizer(tt, return_tensors='pt', padding=True)
yy = tokenizer(tt2, return_tensors='pt', padding=True)

Run a cache step,

gc(xx, yy, reduction='mean')

Here we use reduction='mean' as a loss_kwargs to control loss behavior. With a defined optimizer, the full gradient update can be done as,

optimizer.zero_grad()
gc(xx, yy, reduction='mean')
optimizer.step()

Use Tied Encoder?

This is naturally handled by the (magic of) dynamic graph. You pass shallow copies of the same encoder model to the GradCache init method.

tied_encoder = AutoModel.from_pretrained("bert-base-uncased").cuda()
gc = GradCache(
  models=[tied_encoder , tied_encoder], 
  chunk_sizes=2, 
  loss_fn=loss_fn, 
  get_rep_fn=lambda v: v.pooler_output
)

Under the hood, distinct hooks will be registered to make correct gradient computation.

Distributed Training with Multiple GPUs?

We expect cross process communication of representations to be handled by the loss_fn.

from grad_cache.loss import DistributedContrastiveLoss
loss_fn_dist = DistributedContrastiveLoss()

Properly wrap the the encoder models for gradient reduction,

encoder1_ddp = DistributedDataParallel(
	encoder1, device_ids=[local_rank], output_device=local_rank, find_unused_parameters=True)
encoder2_ddp = DistributedDataParallel(
	encoder2, device_ids=[local_rank], output_device=local_rank, find_unused_parameters=True)

You can initialize the cache use the distributed loss and the DDP models,

gc = GradCache(
  models=[encoder1_ddp, encoder2_ddp], 
  chunk_sizes=2, 
  loss_fn=loss_fn_dist, 
  get_rep_fn=lambda v: v.pooler_output
)

Run a cache step,

gc(xx, yy, no_sync_except_last=True, reduction='mean')

Set no_sync_except_last=True to avoid unnecessary gradient reduction.

Functional Approach

Decorators

If you are developing a new project, we recommend also checking out the decorators we have provided to create higher order functions for cache.

grad_cache.functional.cached(func: Callable[..., Tensor])

A decorator that takes a model call function into a cached compatible version.

func - A function that calls the model and return representation tensor.

Return - A function that returns 1) representation leaf tensors for cache construction, 2) a closure function for the 2nd forward and the cached backward. Call 2) with 1) as argument after calling backward on the loss Tensor.

grad_cache.functional.cat_input_tensor(func: Callable[..., Tensor])

A decorator that concatenates positional and keyword arguments of type List[Tensor] into a single Tensor on the 0th dimension. This can come in handy dealing with results of representation tensors from multiple cached forward.

func - A loss function

Return - Decorated loss function for cached results.

Usage

The functional decorators are particular useful if your data loader is emitting small batches, from which you can construct the big batch. Say you also want to do automatic mixed precision, we first define the model call function and loss function,

from grad_cache.functional import cached, cat_input_tensor

import torch
import torch.nn.functional as F
from torch.cuda.amp import autocast

@cached
@autocast()
def  call_model(model, input):
	return model(**input).pooler_output

@cat_input_tensor
@autocast()
def  contrastive_loss(x, y):
	target = torch.arange(0, y.size(0), int(y.size(0) / x.size(0)), device=x.device)
	scores = torch.matmul(x, y.transpose(0, 1))
	return F.cross_entropy(scores, target=target)

Say you have a DataLoader loader emitting small batches of tuple (xx, yy) of size (M * N) and that you want to train by aggregating 16 small batches to get a batch of (16M * 16N),

cache_x = []
cache_y = []
closures_x = []
closures_y = []

for step, sub_batch in enumerate(loader):  
    xx, yy = sub_batch
    rx, cx = call_model(bert, xx)
    ry, cy = call_model(bert, yy)
    
    cache_x.append(rx)
    cache_y.append(ry)
    closuresx.append(cx)
    closuresy.append(cy)
    
    if (step + 1) % 16 == 0:
        loss = contrastive_loss(cache_x, cache_y)
        scaler.scale(loss).backward()
        
	for f, r in zip(closuresx, cache_x):
            f(r)
        for f, r in zip(closuresy, cache_y):
            f(r)

        cache_x = []
        cache_y = []
        closures_x = []
        closures_y = []
	
        scaler.step(optimizer)
        scaler.update()
        optimizer.zero_grad()

Code Structure

grad_cache/grad_cache.py - Define the GradCache class. The code is under 300 lines including comments. For development, we encourage you to read through it.

grad_cache/functional.py - Define decorators to create higher order function for gradient caching from ordinary model call functions and loss functions.

Owner
Luyu Gao
NLP Research [email protected], CMU
Luyu Gao
The trained model and denoising example for paper : Cardiopulmonary Auscultation Enhancement with a Two-Stage Noise Cancellation Approach

The trained model and denoising example for paper : Cardiopulmonary Auscultation Enhancement with a Two-Stage Noise Cancellation Approach

ycj_project 1 Jan 18, 2022
Official Tensorflow implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation (ICLR 2020)

U-GAT-IT — Official TensorFlow Implementation (ICLR 2020) : Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization fo

Junho Kim 6.2k Jan 04, 2023
Segmentation for medical image.

EfficientSegmentation Introduction EfficientSegmentation is an open source, PyTorch-based segmentation framework for 3D medical image. Features A whol

68 Nov 28, 2022
Catalyst.Detection

Accelerated DL R&D PyTorch framework for Deep Learning research and development. It was developed with a focus on reproducibility, fast experimentatio

Catalyst-Team 12 Oct 25, 2021
Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models (published in ICLR2018)

Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models Pouya Samangouei*, Maya Kabkab*, Rama Chellappa [*: authors co

Maya Kabkab 212 Dec 07, 2022
Omniverse sample scripts - A guide for developing with Python scripts on NVIDIA Ominverse

Omniverse sample scripts ここでは、NVIDIA Omniverse ( https://www.nvidia.com/ja-jp/om

ft-lab (Yutaka Yoshisaka) 37 Nov 17, 2022
Pytorch implemenation of Stochastic Multi-Label Image-to-image Translation (SMIT)

SMIT: Stochastic Multi-Label Image-to-image Translation This repository provides a PyTorch implementation of SMIT. SMIT can stochastically translate a

Biomedical Computer Vision Group @ Uniandes 37 Mar 01, 2022
General-purpose program synthesiser

DeepSynth General-purpose program synthesiser. This is the repository for the code of the paper "Scaling Neural Program Synthesis with Distribution-ba

Nathanaël Fijalkow 24 Oct 23, 2022
Ego4d dataset repository. Download the dataset, visualize, extract features & example usage of the dataset

Ego4D EGO4D is the world's largest egocentric (first person) video ML dataset and benchmark suite, with 3,600 hrs (and counting) of densely narrated v

Meta Research 118 Jan 07, 2023
Implementation of Rotary Embeddings, from the Roformer paper, in Pytorch

Rotary Embeddings - Pytorch A standalone library for adding rotary embeddings to transformers in Pytorch, following its success as relative positional

Phil Wang 110 Dec 30, 2022
Applying PVT to Semantic Segmentation

Applying PVT to Semantic Segmentation Here, we take MMSegmentation v0.13.0 as an example, applying PVTv2 to SemanticFPN. For details see Pyramid Visio

35 Nov 30, 2022
VR-Caps: A Virtual Environment for Active Capsule Endoscopy

VR-Caps: A Virtual Environment for Capsule Endoscopy Overview We introduce a virtual active capsule endoscopy environment developed in Unity that prov

DeepMIA Lab 90 Dec 27, 2022
DABO: Data Augmentation with Bilevel Optimization

DABO: Data Augmentation with Bilevel Optimization [Paper] The goal is to automatically learn an efficient data augmentation regime for image classific

ElementAI 24 Aug 12, 2022
Cooperative Driving Dataset: a dataset for multi-agent driving scenarios

Cooperative Driving Dataset (CODD) The Cooperative Driving dataset is a synthetic dataset generated using CARLA that contains lidar data from multiple

Eduardo Henrique Arnold 124 Dec 28, 2022
Auto White-Balance Correction for Mixed-Illuminant Scenes

Auto White-Balance Correction for Mixed-Illuminant Scenes Mahmoud Afifi, Marcus A. Brubaker, and Michael S. Brown York University Video Reference code

Mahmoud Afifi 47 Nov 26, 2022
Replication of Pix2Seq with Pretrained Model

Pretrained-Pix2Seq We provide the pre-trained model of Pix2Seq. This version contains new data augmentation. The model is trained for 300 epochs and c

peng gao 51 Nov 22, 2022
An essential implementation of BYOL in PyTorch + PyTorch Lightning

Essential BYOL A simple and complete implementation of Bootstrap your own latent: A new approach to self-supervised Learning in PyTorch + PyTorch Ligh

Enrico Fini 48 Sep 27, 2022
An Implementation of SiameseRPN with Feature Pyramid Networks

SiameseRPN with FPN This project is mainly based on HelloRicky123/Siamese-RPN. What I've done is just add a Feature Pyramid Network method to the orig

3 Apr 16, 2022
This is the implementation of the paper "Self-supervised Outdoor Scene Relighting"

Self-supervised Outdoor Scene Relighting This is the implementation of the paper "Self-supervised Outdoor Scene Relighting". The model is implemented

Ye Yu 24 Dec 17, 2022
Official code for article "Expression is enough: Improving traffic signal control with advanced traffic state representation"

1 Introduction Official code for article "Expression is enough: Improving traffic signal control with advanced traffic state representation". The code s

Liang Zhang 10 Dec 10, 2022