Unverified50% confidenceFactExact time
Meta于2023年发布Segment Anything图像分割基础模型，训练了超过11亿个掩码，实现zero-shot通用分割能力
1
Sources
50%
Confidence
Long-term
Relevance
8/13/2026
First Seen
Sources
Related Claims
UnverifiedSegment Anything于2023年发布，训练了超过11亿个掩码，实现了zero-shot通用图像分割能力79% similarUnverifiedMeta的LLaMA于2023年开源后，Hugging Face平台衍生的微调模型在数月内突破数千个70% similarVerifiedMeta released the Llama series of models in 2023, launching the open-source movement for large AI models.69% similarUnverifiedMeta计划在2025年底前建成拥有超过130万块GPU的训练基础设施66% similarVerifiedAfter 2022, diffusion models comprehensively surpassed GANs (Generative Adversarial Networks), which had dominated for many years, in image generation quality, becoming the mainstream paradigm of generative AI.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/742236API
curl https://kongchang.com/api/v1/knowledge/claims/742236MCP
get_claim(id=742236)