NVIDIA releases NVFP4 quantized Qwen-Image with 12B-size weights on Hugging Face
TheZachMueller · x · 2026-10-13
NVIDIA released Qwen-Image-NVFP4 on Hugging Face: an NVFP4 quantization of Qwen's 20.43B-parameter text-to-image diffusion transformer (28.85B full pipeline), produced with Model Optimizer, Apache 2.0 licensed, and deployable via Diffusers with the weights at roughly 12B-parameter size.
More from Infra
- Perplexity's Orbit Event in NYC on Nov 9 Targets Infra and Platform Builders — andrewgwils · 2026-10-13
- Lasers sold out until 2029 as Soitec controls 95% of silicon photonics wafers — demian_ai · 2026-10-13
- Open-source inference project llm-d to share scaling news at PyTorch Conference — PyTorch · 2026-10-13
- Orbital data centers powered by solar energy could ease AI's power crunch — DavidLinthicum · 2026-10-13
- OCP Summit's 20-Minute Sessions Across Three Floors Create Hallway Chaos — firstadopter · 2026-10-13
- PageIndex SDK v0.4 claims tree indexing now cheaper than vector indexing by ~15% — solyarisoftware · 2026-10-13