Pinned Loading

1. StableDiffusion StableDiffusion Public

   A Stable Diffusion 1.x-class latent diffusion model trained from scratch on 2× RTX 5090 (Blackwell) GPUs. Full UNet (~860M params), DDPM/DDIM, LAION pipeline, DDP+BF16.

   Python

2. FusionLLM FusionLLM Public

   Hybrid LLM pre-training framework fusing Multi-Head Latent Attention, Gated Delta Net, DeepSeek MoE, and Multi-Token Prediction — 415M active params, 8.31B tokens, single A100 80GB.

   Python