Search Results for "disk speed test"

...The model accepts combinations of images, camera intrinsics and poses, or even depth cues, then reconstructs consistent 3D geometry suitable for downstream rendering or editing. The pipeline emphasizes both speed and flexibility so creators can go from casual captures to assets without elaborate capture rigs. Outputs can include point clouds, estimated camera parameters, and other 3D representations that plug into typical graphics workflows. The project sits within a broader family of Hunyuan models that explore world generation and 3D-consistent understanding, and this mirror variant makes the reconstruction stack easier to test. ...

...Built on the Gemma 4 26B A4B Mixture-of-Experts architecture, it has 25.2B total parameters and 3.8B active parameters, balancing capability with efficient inference. Its diffusion-based generation produces tokens in parallel 256-token blocks, enabling very high-speed output, with reported generation above 1,100 tokens per second on NVIDIA Hopper H100 in FP8. The model supports a 256K-token context window, configurable thinking mode, native function calling, structured JSON output, and multilingual inference across 35+ languages. The NVFP4 quantization reduces weights and activations from 16-bit to 4-bit, lowering disk size and GPU memory needs for vLLM deployment.