360CVGroup/FG-CLIP — New generation of CLIP with strong fine grained discrimination capability, ICML2026 and ICML2025

Description and use

The **FG‑CLIP** project presents a next‑generation CLIP model that delivers markedly improved fine‑grained discrimination across visual and textual modalities, achieving state‑of‑the‑art performance on benchmark tasks reported at ICML 2025 and ICML 2026. Built as an open‑source library, it offers pretrained weights, easy‑to‑use training scripts, and integration tools for developers and researchers seeking high‑precision multimodal embeddings.

Problem

360CVGroup/FG-CLIP helps make internal knowledge searchable and usable by assistants.

Outcome

Early or unclear: treat as research material until maintenance, releases, docs, and issue activity are verified.

Workflow

• index knowledge bases
• improve search over documents
• ground assistant answers

Evidence

New generation of CLIP with strong fine grained discrimination capability, ICML2026 and ICML2025

Source