๐ฑ ๐ cloneofsimo@gmail.com
lora. Using Low-rank adaptation to quickly fine-tune diffusion models.
7.5kminDiffusion. Self-contained, minimalistic implementation of diffusion models with Pytorch.
1.2kpaint-with-words-sd. Implementation of Paint-with-words with Stable Diffusion : method from eDiff-I that let you generate image from text-labeled segmentation map.
646minRF. Minimal implementation of scalable rectified flow transformers, based on SD3's approach
638minSDXL. Huggingface-compatible SDXL Unet implementation that is readily hackable
439vqgan-training. Train VAE like a boss
313d3pm. Minimal Implementation of a D3PM in pytorch
306consistency_models. Unofficial Implementation of Consistency Models in pytorch
259min-max-gpt. Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training
132realformer-pytorch. Implementation of RealFormer using pytorch
101magicmix. Unofficial Implementation of MagicMix
98scaling-guide. WIP
96min-fsdp. Python
93ezmup. Simple implementation of muP, based on Spectral Condition for Feature Learning. The implementation is SGD only, dont use it for Adam
88t2i-adapter-diffusers. Python
86sdxl_inversions. Jupyter Notebook
85promptplusplus. Jupyter Notebook
73ptx-tutorial-by-aislop. PTX-Tutorial Written Purely By AIs (Deep Research of Openai and Claude 3.7)
66sd-various-ideas. Jupyter Notebook
56karras-power-ema-tutorial. Python
53insightful-nn-papers. These papers will provide unique insightful concepts that will broaden your perspective on neural networks and deep learning
48clipping-CLIP-to-GAN. Python
42imagenet.int8. Python
40zeroshampoo. Python
33fim-llama-deepspeed. Python
32repa-rf. Python
32infinite-fractal-stream. Jupyter Notebook
30minSAE. Python
30auto_llm_codebase_analysis. Python
27min-max-in-dit. Python
27minVJEPA. Python
25project_RF. Python
24minDinoV2. Python
24efae. Python
24inversion_edits. Jupyter Notebook
21planning-with-diffusion-tutorial. Jupyter Notebook
18zeroshot-storytelling. Github repository for Zero Shot Visual Storytelling
15blocked-decorr. Python
13ptar. C++
13poly2SOP. Transformer takes a polynomial, expresses it as sum of powers.
11minMomentMatching. Python
11smallest_working_performer. Python
10minmoe. Python
6n-body-dynamic-cuda. Cuda
6smallest_working_gpt. gpt that is even smaller
6rectified-flow. Jupyter Notebook
5torchcu. Python
5reverse_eng_deepspeed_study. DeepSpeed Study, focused on reverse engineering and enhancing documentation
5imgdataset_process. Python
4lora_dreambooth_replicate. Jupyter Notebook
4SemanticSegmentationTrainerTemplate. Python
1cv. Simple CV (pdf, Latex)
1netfilter. C
1unn-lstm-torch. Python
1Super-Simple-LSTM-Template. Python
1send-arp. C++
1Freshman_2. Lecture notes from my Freshman 2nd semester
1