Training Pipelines

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型，实现了包括增量预训练(PT)、有监督微调(SFT)、RLHF、DPO、ORPO、GRPO。

A procedural Blender pipeline for photorealistic training image generation

A bidirectional pipeline parallelism algorithm for computation-communication overlap in DeepSeek V3/R1 training.

A pipeline parallel training script for diffusion models.

A nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.

Real-time multi-person action recognition from OpenPose skeleton sequences, with tracking, feature extraction, model training, and inference pipelines for research and prototyping.

A minimalist SOTA LaTeX OCR model with only 20M parameters, running in browser. Full training pipeline available for self-reproduction. | 超轻量SOTA LaTeX公式识别模型，仅20M参数量，可在浏览器中运行。训练全流程代码开源，以便自学复现。