HF guide: run any new model in Diffusers via Modular custom blocks, no official PR needed
RisingSayak · x · 2026-11-17
Hugging Face community author OzzyGT published a guide on integrating new models with Modular Diffusers without waiting for official support. The demo is IDOL (single image to 3D Gaussian splat): convert weights and custom blocks to two Hub repos, then load with ModularPipeline.frompretrained(..., trustremotecode=True) in a few lines, inheriting Diffusers' offloading, torch.compile, and attention backends. The pipeline returns Gaussian tensors ready to render or save as .ply, plus camera pose. Author used Claude Code with Opus 5.5 plus skills to do the conversion — Astra failed at the same task. Caveat: only load repos you trust.
Related event: Hugging Face guide shows how to run new models before official integration(2 posts)→
More from coding & agent
- Codex lacks visual indicator for multi-day agent runs, user resorts to renaming tasks — chrisalbon · 2026-11-17
- Poolside's 'everything collapses into the model' thesis shapes new agent runtime Rex — sull · 2026-11-17
- FrontierCS 2 benchmark challenges AI agents to beat published CS research blind — arena · 2026-11-17
- VPS MCP Server lets AI models manage Docker containers via Portainer API — modelcontextprotocol · 2026-11-17
- Qualcomm demos an on-device decision model routing AI requests in 8.6ms on NPU — carrycooldude · 2026-11-17
- User's open letter to OpenAI: Dot agent degraded after a week, switching to Claude Duet — Danny7Guns · 2026-11-17