AI Model2024
Generates images from text prompts using a 12-billion-parameter rectified-flow transformer trained with guidance distillation for more efficient sampling. Distributed with diffusers/ComfyUI support and multiple conditioning/editing variants; weights released under a non-commercial license.
AI Audio2024
Chains four swappable open modules — voice activity detection, speech-to-text, an LLM, and text-to-speech — into a local voice agent that needs no proprietary APIs. Runs on CUDA, Apple Silicon, or Docker, with an OpenAI-compatible realtime WebSocket mode.
AI Train2024
Trains a sub-100M-parameter LLM from scratch — pretraining, SFT, LoRA, DPO/RLHF, and distillation, sized from ~26M up to ~100M-plus dense and MoE. Headline figure: the ~64M minimind-3 variant's SFT stage runs 1 epoch in ~2h and ~3 RMB on one NVIDIA 3090.
AI Deploy2024
Runs and optimizes ML and generative-AI models on-device across mobile, desktop, web, and IoT. Successor to TensorFlow Lite, it adds automated GPU/NPU accelerator selection and zero-copy buffer interop to cut latency without cloud round-trips.
AI Image2024
Trains a 65M-parameter vision-language model from scratch in ~2 hours on one RTX 3090, about 3 RMB (~$0.40) of GPU rental. Connects a frozen SigLIP2 encoder to a small MiniMind LLM via a two-layer MLP projector; full PyTorch code for pretraining and SFT.
Embodied AI2024
GPU‑accelerated framework for training physically simulated humanoid characters and robots using reinforcement learning and motion imitation. Provides a modular multi‑backend simulator stack, large‑scale multi‑GPU training recipes, built‑in motion retargeting and an ONNX deployment pathway to real robots.
AI Infra2024
A compact domain-specific language for writing high-performance GPU/CPU kernels (GEMM, FlashAttention, sparse kernels) with Python-like syntax. It provides tiling/pipelining primitives, a TVM-based compiler and multiple backends (CUDA/CuTeDSL, NVRTC, WebGPU, Metal, Ascend) for operator-level performance work.
AI Model2024
High-resolution image and video generation codebase and models that run with far lower compute and memory than typical diffusion systems. Uses linear-attention DiT variants, aggressive latent compression, and inference-scaling to support text-to-image (up to 4K), fast one/few-step generation, and efficient video pipelines.
AI Train2024
Open-source HybridFlow implementation for RL post-training of LLMs. Decouples control flow from compute so PPO, GRPO, GSPO and DAPO share one dataflow; pairs FSDP/Megatron with vLLM/SGLang rollout and reports 1.5-20x throughput over prior RLHF stacks.
A GitHub repository of learning notes and code dedicated to ML + SYS (machine learning systems). It collects tutorials, code walkthroughs and engineering notes on RLHF, distributed training (FSDP, Megatron), inference and scheduling (SGLang, vllm), quantization, CUDA/GPU optimization, system design, and practical engineering.
AI Video2024
Generates video from text or images via a DiT-based latent diffusion model: text-to-video, image-to-video, frame extension, and multi-keyframe conditioning in one model. A distilled 2B variant runs near real-time on one H100; 13B for higher quality.
AI Model2024
Generates high-quality, editable 3D assets from text or images and decodes to radiance fields, 3D Gaussians, or textured meshes. Ships pretrained models up to 2B parameters, a 500K asset dataset and training code; best used with image conditioning and a ≥16GB NVIDIA GPU.