Search results

Where does the reduction of TTFT comes from in LLaVa-UHD v3? Is it from ViT-UHD infer speed or visual token reduction in prefilling of LLM?

Status: Open.

#37 In thunlp/LLaVA-UHD;

· lawsonxwl opened on Feb 5, 2026

v2

Status: Open.

#36 In thunlp/LLaVA-UHD;

· zzzzzzyc opened on Jan 26, 2026

数据样例问题

Status: Open.

#35 In thunlp/LLaVA-UHD;

· yhl41001 opened on Dec 30, 2025

Trained Resample with Siglip Got inconvergence loss

Status: Open.

#22 In thunlp/LLaVA-UHD;

· lucasjinreal opened on Apr 23, 2024

Loss Convergence and whether ViT is Trained

Status: Open.

#18 In thunlp/LLaVA-UHD;

· SuperStacie opened on Apr 17, 2024

[Question] Proof about Range of Slice Aspect Ratios

Status: Open.

#15 In thunlp/LLaVA-UHD;

· JJJYmmm opened on Apr 17, 2024

Comparing with LLaVA 1.6 Next

Status: Open.

#1 In thunlp/LLaVA-UHD;

· choyakawa opened on Mar 19, 2024