This PyTorch package implements MoEBERT: from BERT to Mixture-of-Experts via Importance-Guided Adaptation (NAACL 2022).

Related tags

Deep LearningMoEBERT
Overview

MoEBERT

This PyTorch package implements MoEBERT: from BERT to Mixture-of-Experts via Importance-Guided Adaptation (NAACL 2022).

Installation

  • Create and activate conda environment.
conda env create -f environment.yml
  • Install Transformers locally.
pip install -e .
  • Note: The code is adapted from this codebase. Arguments regarding LoRA and adapter can be safely ignored.

Instructions

MoEBERT targets task-specific distillation. Before running any distillation code, a pre-trained BERT model should be fine-tuned on the target task. Path to the fine-tuned model should be passed to --model_name_or_path.

Importance Score Computation

  • Use bert_base_mnli_example.sh to compute the importance scores, add a --preprocess_importance argument, remove the --do_train argument.
  • If multiple GPUs are used to compute the importance scores, a importance_[rank].pkl file will be saved for each GPU. Use merge_importance.py to merge these files.
  • To use the pre-computed importance scores, pass the file name to --moebert_load_importance.

Knowledge Distillation

  • For GLUE tasks, see examples/text-classification/run_glue.py.
  • For question answering tasks, see examples/question-answering/run_qa.py.
  • Run bash bert_base_mnli_example.sh as an example.
  • The codebase supports different routing strategies: gate-token, gate-sentence, hash-random and hash-balance. Choices should be passed to --moebert_route_method.
    • To use hash-balance, a balanced hash list needs to be pre-computed using hash_balance.py. Path to the saved hash list should be passed to --moebert_route_hash_list.
    • Add a load balancing loss by setting --moebert_load_balance when using trainable gating mechanisms.
    • The sentence-based gating mechanism (gate-sentence) is advantageous for inference because it induces significantly less communication overhead compared with token-level routing methods.
Owner
Simiao Zuo
PhD Student @ Georgia Tech
Simiao Zuo
MADE (Masked Autoencoder Density Estimation) implementation in PyTorch

pytorch-made This code is an implementation of "Masked AutoEncoder for Density Estimation" by Germain et al., 2015. The core idea is that you can turn

Andrej 498 Dec 30, 2022
clDice - a Novel Topology-Preserving Loss Function for Tubular Structure Segmentation

README clDice - a Novel Topology-Preserving Loss Function for Tubular Structure Segmentation CVPR 2021 Authors: Suprosanna Shit and Johannes C. Paetzo

110 Dec 29, 2022
Code for ECCV 2020 paper "Contacts and Human Dynamics from Monocular Video".

Contact and Human Dynamics from Monocular Video This is the official implementation for the ECCV 2020 spotlight paper by Davis Rempe, Leonidas J. Guib

Davis Rempe 207 Jan 05, 2023
A platform for intelligent agent learning based on a 3D open-world FPS game developed by Inspir.AI.

Wilderness Scavenger: 3D Open-World FPS Game AI Challenge This is a platform for intelligent agent learning based on a 3D open-world FPS game develope

46 Nov 24, 2022
Implementation of "Scaled-YOLOv4: Scaling Cross Stage Partial Network" using PyTorch framwork.

YOLOv4-large This is the implementation of "Scaled-YOLOv4: Scaling Cross Stage Partial Network" using PyTorch framwork. YOLOv4-CSP YOLOv4-tiny YOLOv4-

Kin-Yiu, Wong 2k Jan 02, 2023
MLOps will help you to understand how to build a Continuous Integration and Continuous Delivery pipeline for an ML/AI project.

page_type languages products description sample python azure azure-machine-learning-service azure-devops Code which demonstrates how to set up and ope

1 Nov 01, 2021
Dynamic Realtime Animation Control

Our project is targeted at making an application that dynamically detects the user’s expressions and gestures and projects it onto an animation software which then renders a 2D/3D animation realtime

Harsh Avinash 10 Aug 01, 2022
ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training

ActNN : Activation Compressed Training This is the official project repository for ActNN: Reducing Training Memory Footprint via 2-Bit Activation Comp

UC Berkeley RISE 178 Jan 05, 2023
Pytorch implementation for Patient Knowledge Distillation for BERT Model Compression

Patient Knowledge Distillation for BERT Model Compression Knowledge distillation for BERT model Installation Run command below to install the environm

Siqi 180 Dec 19, 2022
Differentiable Neural Computers, Sparse Access Memory and Sparse Differentiable Neural Computers, for Pytorch

Differentiable Neural Computers and family, for Pytorch Includes: Differentiable Neural Computers (DNC) Sparse Access Memory (SAM) Sparse Differentiab

ixaxaar 302 Dec 14, 2022
Project page for our ICCV 2021 paper "The Way to my Heart is through Contrastive Learning"

The Way to my Heart is through Contrastive Learning: Remote Photoplethysmography from Unlabelled Video This is the official project page of our ICCV 2

36 Jan 06, 2023
Deep Learning and Logical Reasoning from Data and Knowledge

Logic Tensor Networks (LTN) Logic Tensor Network (LTN) is a neurosymbolic framework that supports querying, learning and reasoning with both rich data

171 Dec 29, 2022
Parameterising Simulated Annealing for the Travelling Salesman Problem

Parameterising Simulated Annealing for the Travelling Salesman Problem

Gary Sun 55 Jun 15, 2022
Official implementation of Few-Shot and Continual Learning with Attentive Independent Mechanisms

Few-Shot and Continual Learning with Attentive Independent Mechanisms This repository is the official implementation of Few-Shot and Continual Learnin

Chikan_Huang 25 Dec 08, 2022
As-ViT: Auto-scaling Vision Transformers without Training

As-ViT: Auto-scaling Vision Transformers without Training [PDF] Wuyang Chen, Wei Huang, Xianzhi Du, Xiaodan Song, Zhangyang Wang, Denny Zhou In ICLR 2

VITA 68 Sep 05, 2022
Official Keras Implementation for UNet++ in IEEE Transactions on Medical Imaging and DLMIA 2018

UNet++: A Nested U-Net Architecture for Medical Image Segmentation UNet++ is a new general purpose image segmentation architecture for more accurate i

Zongwei Zhou 1.8k Jan 07, 2023
Relaxed-machines - explorations in neuro-symbolic differentiable interpreters

Relaxed Machines Explorations in neuro-symbolic differentiable interpreters. Baby steps: inc_stop Libraries JAX Haiku Optax Resources Chapter 3 (∂4: A

Nada Amin 6 Feb 02, 2022
UDP++ (ECCVW 2020 Oral), (Winner of COCO 2020 Keypoint Challenge).

UDP-Pose This is the pytorch implementation for UDP++, which won the Fisrt place in COCO Keypoint Challenge at ECCV 2020 Workshop. Top-Down Results on

20 Jul 29, 2022
Official implementation for CVPR 2021 paper: Adaptive Class Suppression Loss for Long-Tail Object Detection

Adaptive Class Suppression Loss for Long-Tail Object Detection This repo is the official implementation for CVPR 2021 paper: Adaptive Class Suppressio

CASIA-IVA-Lab 67 Dec 04, 2022
The source code for the Cutoff data augmentation approach proposed in this paper: "A Simple but Tough-to-Beat Data Augmentation Approach for Natural Language Understanding and Generation".

Cutoff: A Simple Data Augmentation Approach for Natural Language This repository contains source code necessary to reproduce the results presented in

Dinghan Shen 49 Dec 22, 2022