GLU Variants Improve Transformer

BlinkDL/RWKV-LM 12 Feb 2020

Gated Linear Units (arXiv:1612. 08083) consist of the component-wise product of two linear projections, one of which is first passed through a sigmoid function.

330
1.00 stars / hour

OPT: Open Pre-trained Transformer Language Models

facebookresearch/metaseq 2 May 2022

Large language models, which are often trained for hundreds of thousands of compute days, have shown remarkable capabilities for zero- and few-shot learning.

Hate Speech Detection Language Modelling +1

2,824
0.87 stars / hour

A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation

spotify/basic-pitch 18 Mar 2022

Despite its simplicity, benchmark results show our system's note estimation to be substantially better than a comparable baseline, and its frame-level accuracy to be only marginally below those of specialized state-of-the-art AMT systems.

Event Detection Frame +1

72
0.83 stars / hour

Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning

r-three/t-few 11 May 2022

ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made.

Few-Shot Text Classification

74
0.69 stars / hour

SVTR: Scene Text Recognition with a Single Visual Model

PaddlePaddle/PaddleOCR 30 Apr 2022

Dominant scene text recognition models commonly contain two building blocks, a visual model for feature extraction and a sequence model for text transcription.

Scene Text Recognition

21,760
0.61 stars / hour

DualDis: Dual-Branch Disentangling with Adversarial Learning

ThomasRobertFr/deep-learning-figures 3 Jun 2019

To effectively separate the information, we propose to use a combination of regular and adversarial classifiers to guide the two branches in specializing for class and attribute information respectively.

Data Augmentation Image Manipulation +1

54
0.45 stars / hour

Surface Representation for Point Clouds

hancyran/RepSurf 11 May 2022

Based on a simple baseline of PointNet++ (SSG version), Umbrella RepSurf surpasses the previous state-of-the-art by a large margin for classification, segmentation and detection on various benchmarks in terms of performance and efficiency.

3D Object Detection 3D Point Cloud Classification +2

33
0.39 stars / hour

View Synthesis with Sculpted Neural Points

princeton-vl/snp 12 May 2022

We address the task of view synthesis, which can be posed as recovering a rendering function that renders new views from a set of existing images.

26
0.36 stars / hour

READ: Large-Scale Neural Scene Rendering for Autonomous Driving

JOP-Lee/READ-Large-Scale-Neural-Scene-Rendering-for-Autonomous-Driving 11 May 2022

In this paper, a large-scale neural rendering method is proposed to synthesize the autonomous driving scene~(READ), which makes it possible to synthesize large-scale driving scenarios on a PC through a variety of sampling schemes.

3D Scene Reconstruction Autonomous Driving +4

37
0.35 stars / hour

Implementation of an Automated Learning System for Non-experts

industryessentials/ymir 26 Mar 2022

This paper detailed the engineering system implementation of an automated machine learning system called YMIR, which completely relies on graphical interface to interact with users.

131
0.34 stars / hour