GigaPath-Flash and GigaTIME-Flash: Efficient Pathology Foundation Models for Whole-Slide and Tumor Microenvironment Analysis
The paper presents smaller open-weight pathology models that keep most slide-level performance while sharply reducing compute.
GigaPath-Flash pairs a 22M-parameter ViT-S tile encoder with a 21M-parameter LongNet slide encoder, distilled from the larger GigaPath model. The authors say it retains 97% of GigaPath’s average slide-level performance with 50x less compute. GigaTIME-Flash uses the same backbone to predict tumor immune microenvironment signals from H&E images, outperforming the earlier CNN-based GigaTIME while running 6x faster and using 8x less GPU memory. The model family is released as open weights under Apache 2.0. ArXiv · AI/CL/LG's note
GigaPath-Flash pairs a 22M-parameter ViT-S tile encoder with a 21M-parameter LongNet slide encoder, distilled from the larger GigaPath model. The authors say it retains 97% of GigaPath’s average slide-level performance with 50x less compute. GigaTIME-Flash uses the same backbone to predict tumor immune microenvironment signals from H&E images, outperforming the earlier CNN-based GigaTIME while running 6x faster and using 8x less GPU memory. The model family is released as open weights under Apache 2.0. ArXiv · AI/CL/LG's note
score 6