nvidia/Nemotron-Labs-TwoTower-30B-A3B-Base-BF16

Upload paper.pdf

add -Labs to model name

Update to Nemotron-Labs-TwoTower

add arXiv URL

relax figure

reax figure

add comparison figure

Upload table_comparison.png with huggingface_hub

rename backbone annotation

Upload paper.pdf with huggingface_hub

upload paper until arxiv clears

add scores

refine readme

Update model card (README) and tidy inference scaffolding

Fix NaN corruption in long-context diffusion (fp32 denoiser SSM scan) + multi-request inference

Add cached generate_ar (ST-AR baseline) + single-step AR/mock-AR context extend (stock parity)

faster inferneece

Two-tower mask-diffusion weights (corrected conversion, iter_0085937)

Two-tower mask diffusion: fix denoiser (adaLN norm order, bidirectional in-block attention, block-wise chunk-scan Mamba) + fp64 router; refresh README

Upload folder using huggingface_hub

Upload folder using huggingface_hub

initial commit