kimrass / detr

PyTorch implementation of 'DETR' (Carion et al., 2020) from scratch.

1. Resume

RESUME

2. 'Baekjoon Online Judge' Solved Rank

3. PyTorch Implementations From Scratch

Vision
2014	VAE	Kingma and Welling	[✓] Training on MNIST [✓] Encoder output visualization [✓] Decoder output visualization
2015	CAM	Zhou et al.	[✓] Application to GoogleNet
2016	Gatys et al., 2016	Gatys et al.	[✓] Application to VGGNet-19
	YOLO	Redmon et al.	[✗] Training on VOC 2012 [✗] Class probability map [✗] Ground truth vlisualization on grid
	DCGAN	Radford et al.	[✓] Training on CelebA at 64 × 64 [✓] Sampling [✓] Latent space interpolation
	Noroozi et al., 2016	Noroozi et al.	[✓] Architecture [✓] Chromatic aberration [✓] Permutation set
	Zhang et al., 2016	Zhang et al.	[✓] Empirical probability distribution [✗] Color space
2014 2017	Conditional GAN WGAN-GP	Mirza et al. Gulrajani et al.	[✓] Training on MNIST
2016 2017	PixelCNN VQ-VAE	Oord et al. Oord et al.	[✓] Training on Fashion MNIST [✓] Training on CIFAR-10
2017	Pix2Pix	Isola et al.	[✓] Training on Google Maps [✓] Training on Facades [✗] Inference on larger resolution
	CycleGAN	Zhu et al.	[✓] Training on Monet to photo [✓] Training on Vangogh to photo [✓] Training on Cezanne to photo [✓] Training on Ukiyo-e to photo [✓] Training on Horse to zebra [✓] Training on Summer to winter
	Noroozi et al., 2017	Noroozi et al.	[✓] Constrastive loss
2018	PGGAN	Karras et al.	[✓] Training on CelebA-HQ at 512 × 512
	DeepLab v3	Chen et al.	[✓] Training on VOC 2012 [✓] Prediction on VOC 2012 validation set [✓] Average mIoU
	PixelLink	Deng et al.	[✓] Architecture [✓] Instance-balanced cross entropy loss [✓] Post-processing
	RotNet	Gidaris et al	[✓] Attention map visualization
2020	STEFANN	Roy et al.	[✓] FANnet architecture [✓]Training FANnet on Google Fonts [✓] Custom Google Fonts dataset [✓] Average SSIM
	DDPM	Ho et al.	[✓] Training on CelebA at 32 × 32 [✓] Training on CelebA at 64 × 64 [✓] Denoising process visualization [✓] Linear interpolation sampling [✓] Coarse-to-fine sampling
	DDIM	Song et al.	[✓] Sampling [✓] Spherical interpolation sampling [✓] Interpolation on grid sampling [✓] Truncated normal
	ViT	Dosovitskiy et al.	[✓] Training on CIFAR-10 [✓] Training on CIFAR-100 [✓] Attention Roll-out [✓] Position embedding similarity [✓] Position embedding interpolation Extra [✓] CutOut [✓] Hide-and-Seek [✓] CutMix
	SimCLR	Chen et al.	[✓] Normalized temperature-scaled cross entropy loss [✓] Data augmentation [✓] Pixel intensity histogram
	DETR	Carion et al.	[✓] Architecture [✗] Batch normalization freezing [✗] Data preparation [✗] Training on COCO 2017
2021	Improved DDPM	Nichol and Dhariwal	[✓] Cosine diffusion schedule
	Classifier-Guidance	Dhariwal and Nichol	[✗] AdaGN [✗] BiGGAN Upsample/Downsample [✗] Improved DDPM sampling [✗] Conditional/Unconditional models [✗] Super-resolution model [✗] Interpolation
	ILVR	Choi et al.	[✓] Sampling from single reference [✓] Sampling from various scale factors [✓] Sampling from various conditioning range
	SDEdit	Meng et al.	[✓] User input stroke simulation
	MAE	He et al.	[✓] MAE architecture for pre-training [✗] MAE architecture for self-supervised learning [✗] Training on ImageNet-1K [✗] Fine-tuning [✗] Linear probing
	Copy-Paste	Ghiasi et al.	[✓] Large scale jittering [✓] Copy-Paste (within mini-batch) [✗] Gaussian filter
	ViViT	Arnab et al.
2022	CFG	Ho et al.
Language
2017	Transformer	Vaswani et al.	[✓] Architecture [✓] Position encoding visualization
2019	BERT	Devlin et al.	[✓] BookCorpus data pre-processing [✓] Architecture [✓] Masked language modeling [✓] SQuAD data pre-processing [✓]SWAG data pre-processing
	Sentence-BERT	Reimers et al.	[✓] Classification loss [✓] Regression loss [✓] Constrastive loss [✓] STSb data pre-processing [✓] WikiSection data pre-processing [✗] NLI data pre-processing
	RoBERTa	Liu et al.	[✓] BookCorpus data pre-processing [✓] Masked language modeling [✗] BookCorpus data pre-processing (SEGMENT-PAIR + NSP) [✗] BookCorpus data pre-processing (SENTENCE-PAIR + NSP) [✓] BookCorpus data pre-processing (FULL-SENTENCES) [✗] BookCorpus data pre-processing (DOC-SENTENCES)
Vision-Language
2021	CLIP	Radford et al.	[✓] Training on Flickr8k + Flickr30k [✓] Zero-shot classification on ImageNet1k (mini) [✓] Linear classification on ImageNet1k (mini)