qwopqwop200

GPTQ-for-LLaMa. 4 bits quantization of LLaMA using GPTQ

gptqlora. GPTQLoRA: Efficient Finetuning of Quantized LLMs with GPTQ

lama-with-maskdino. automatic image inpainting (lama(with refinement) and maskdino)

ko-arena-hard-auto. Ko-Arena-Hard-Auto: An automatic LLM benchmark for Korean

stable-diffusion-webui-promptgen-danbooru. stable-diffusion-webui-promptgen

GPTQ-for-KoAlpaca. Python

SoftPool. softpool implementation(Refining activation downsampling with SoftPool) This is an unofficial implementation. https://arxiv.org/pdf/2101.00440v2.pdf

AutoQuant. Python

MaxVIT-pytorch. MaxVIT implementation(MaxViT: Multi-Axis Vision Transformer) This is an unofficial implementation. https://arxiv.org/abs/2204.01697

llama-danbooru-qlora. Jupyter Notebook

ALMA-en2ko. This is repository for ALMA translation models.

Neighborhood-Attention-Transformer. NAT implementation(Neighborhood Attention Transformer) This is an unofficial implementation. https://arxiv.org/pdf/2204.07143.pdf

pale-transformer-pytorch. Pale Transformer implementation(Pale Transformer: A General Vision Transformer Backbone with Pale-Shaped Attention) This is an unofficial implementation. https://arxiv.org/abs/2112.14000

NatIR. NatIR: Image Restoration Using Neighborhood-Attention-Transformer

KoLIMA. Jupyter Notebook

MLP-Mixer-tf2. MLP-Mixer implementation(MLP-Mixer: An all-MLP Architecture for Vision) This is an unofficial implementation. https://arxiv.org/pdf/2105.01601v1.pdf

Magneto-pytorch. Magneto implementation(Foundation Transformers) This is an unofficial implementation. https://arxiv.org/abs/2210.06423

Swin-MLP-Mixer. This code is a structure that combines swim-transformer and mlp mixer, and performance may be poor because I didn’t train and test this model

Subtitles-generator-with-whisper. Subtitles generator using whisper and translator

D-Adaptation-Adan. Adan with D-Adaptation automatic step-sizes

transformers-t5. 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.

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

psnr_ssim_ycbcr. Code for calculating psnr and ssim in y channel in ycbcr.This code is based on BasicSR (https://github.com/xinntao/BasicSR).

halonet-tf2. halonet implementation(Scaling Local Self-Attention for Parameter Efficient Visual Backbones) This is an unofficial implementation.https://arxiv.org/pdf/2103.12731v2.pdf