Jonathan Chang

minLoRA. minLoRA: a minimal PyTorch library that allows you to apply LoRA to any PyTorch model.

flex-nano-vllm. FlexAttention based, minimal vllm-style inference engine for fast Gemma 2 inference.

fine-tune-models. Jupyter Notebook

efficient-prompt-to-prompt. Jupyter Notebook

anim_e. Anim·E, Anime Enhanced dalle mini

llmproc. LLMProc: Unix-inspired runtime that treats LLMs as processes.

LoRAnanoGPT. The simplest, fastest repository for training/finetuning medium-sized GPTs.

claude-code-system-prompts. JavaScript

t5-flex-attention. T5 model optimized with FlexAttention

multi_task_utils. A simple python package for multi-task training: wrappers for pytorch DataLoader and pytorch-lightning DataModule

tw_rouge. ROUGE score calculator with traditional chinese word segmentation

forking-an-agent. Python

kodx. Codex-inspired local background agent

mochi-1. The best OSS video generation models

mcp-registry. Python

vFLUX. Python

gRPC-monitor. A minimum viable multi-server resource monitoring web app using gRPC and Plotly Dash

ft5-demo. ft5 demo code

blood-taiwan. Periodically scrape data from Taiwan blood bank with GitHub workflows. Data source: http://www.blood.org.tw/Internet/main/index.aspx

aishell. A transparent shell wrapper that for building context-aware AI tools

agent-chat. Python

voyager-mcp. Python

min-dalle. min(DALL·E) is a fast, minimal port of DALL·E Mini to PyTorch

keras_big_vision. Keras implementation of SigLIP

cookiecutter-data-science. A logical, reasonably standardized, but flexible project structure for doing and sharing data science work.

ft5-demo-space. Mirror of the huggingface demo space

fnet-generation. Python

accelerated-stable-diffusion. Python

soft-tokenizer. soft-tokenizer

wikimcp. JavaScript

nano-vllm. Nano vLLM

consistencydecoder. Consistency Distilled Diff VAE

kaggleupload. An unofficial wrapper around kaggle-api

huggingface_hub. Client library to download and publish models and other files on the huggingface.co hub

llmcp. Python

accelerate. 🚀 A simple way to train and use PyTorch models with multi-GPU, TPU, mixed-precision

swarm2. Educational framework exploring ergonomic, lightweight multi-agent orchestration. Managed by OpenAI Solution team.

diffusers. 🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch

my-line-bot. Python