Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection

Last update: Sep 08, 2022

Overview

Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection (NimPme)

The official implementation of Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection

The code is built on TFA

Requirements

Linux with Python >= 3.6
PyTorch >= 1.4
torchvision that matches the PyTorch installation
CUDA 10.0, 10.1, 10.2
GCC >= 4.9

Getting Started

Evaluation with pre-trainied 10-shot final detecor

we provide the pre-trainied 10-shot final detecor

python3 -m tools.test_net --num-gpus 1 \
        --config-file configs/PascalVOC-detection/split1/faster_rcnn_R_101_FPN_ft_all1_10shot.yaml \
        --eval-only

Training & Evaluation in Command Line

To train a base detector, run

python3 -m tools.train_net --num-gpus 1 \
        --config-file configs/PascalVOC-detection/split1/faster_rcnn_R_101_FPN_base1.yaml

fine-tune the detector with novel set

python3 -m tools.ckpt_surgery \
        --src1 checkpoints/coco/faster_rcnn/faster_rcnn_R_101_FPN_base/model_final.pth \
        --method randinit \
        --save-dir checkpoints/coco/faster_rcnn/faster_rcnn_R_101_FPN_all
        --coco

python3 -m tools.train_net --num-gpus 1 \
        --config-file configs/COCO-detection/faster_rcnn_R_101_FPN_ft_all_1shot.yaml \
        --opts MODEL.WEIGHTS $WEIGHTS_PATH

fine-tune the detector with pseudo set

python3 -m tools.genarate_pseudo --num-gpus 1
python3 -m tools.train_feature --num-gpus 1

Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection

Related tags

Overview

Novel Instances Mining with Pseudo-Margin Evaluation for Few-Shot Object Detection (NimPme)

Getting Started

Evaluation with pre-trainied 10-shot final detecor

Training & Evaluation in Command Line

Owner

A library for researching neural networks compression and acceleration methods.

The code for the CVPR 2021 paper Neural Deformation Graphs, a novel approach for globally-consistent deformation tracking and 3D reconstruction of non-rigid objects.

Source code for our paper "Empathetic Response Generation with State Management"

CompilerGym is a library of easy to use and performant reinforcement learning environments for compiler tasks

MAVE: : A Product Dataset for Multi-source Attribute Value Extraction

ManipNet: Neural Manipulation Synthesis with a Hand-Object Spatial Representation - SIGGRAPH 2021

Text-Based Ideal Points

A simple rest api that classifies pneumonia infection weather it is Normal, Pneumonia Virus or Pneumonia Bacteria from a chest-x-ray image.

Tensorflow-Project-Template - A best practice for tensorflow project template architecture.

Simple torch.nn.module implementation of Alias-Free-GAN style filter and resample

Out-of-Domain Human Mesh Reconstruction via Dynamic Bilevel Online Adaptation

Face Identity Disentanglement via Latent Space Mapping [SIGGRAPH ASIA 2020]

Retinal vessel segmentation based on GT-UNet

ConE: Cone Embeddings for Multi-Hop Reasoning over Knowledge Graphs

Rank 3 : Source code for OPPO 6G Data Generation Challenge

Group Activity Recognition with Clustered Spatial Temporal Transformer

Development kit for MIT Scene Parsing Benchmark

Data and code for the paper "Importance of Kernel Bandwidth in Quantum Machine Learning"

Baseline inference Algorithm for the STOIC2021 challenge.

MLP-Numpy - A simple modular implementation of Multi Layer Perceptron in pure Numpy.