ChandlerIdeaCreator / AIC-Webly-supervised-Fine-Grained-Image-Recognition-National-Second-PrizeView on GitHub
框架核心是两阶段“粗筛-精滤”数据清洗流程。首先,利用CLIP的多门控决策逻辑进行宏观粗筛,精准剔除插画、图表等非摄影类噪声。随后,利用DINOv2的细粒度特征,创新采用“相对边际分数”识别处于类别边界的混淆样本,并结合GMM模型为各类别动态确定清洗标准。整个流程内置最小样本保留机制,以保护长尾分布下的稀有类别。 在清洗后的高纯度数据上,我们采用非对称协同训练策略。通过Swin Transformer与ConvNeXt双网络,利用其结构差异形成互补的错误识别模式,高效实现相互监督与纠错。同时,我们创新地对模型达成共识的样本目标化应用Balanced Softmax Loss,在精准应对长尾分布的同时,避免了放大噪声的影响。
☆207Apr 21, 2026Updated last month
Alternatives and similar repositories for AIC-Webly-supervised-Fine-Grained-Image-Recognition-National-Second-Prize
Users that are interested in AIC-Webly-supervised-Fine-Grained-Image-Recognition-National-Second-Prize are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- 一款基于 SOTA 模型 BiRefNet 开发的高精度 AI 抠图工具☆62Jan 22, 2026Updated 4 months ago
- ☆844Feb 18, 2026Updated 4 months ago
- 🔥minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,minerproxy,矿池抽水,矿池中转,矿场运维专用☆3,658May 22, 2026Updated 3 weeks ago
- Privacy-focused fork of Claude Code with all telemetry and analytics removed☆143Apr 6, 2026Updated 2 months ago
- 数字底座是一款面向大型政府、企业数字化转型,基于身份认证、组织架构、岗位职务、应用系统、资源角色、数据目录、安全控制等功能构建的统一且安全的管理支撑平台。数字底座基于三员管理模式,具备微服务、多租户、容器化和国产化,支持用户利用代码生成器快速构建自己的业务应用,同时可关联诸…☆2,595Updated this week
- Open source password manager - Proton Pass • AdSecurely store, share, and autofill your credentials with Proton Pass, the end-to-end encrypted password manager trusted by millions.
- [CVPR 2026] PromptEnhancer is a prompt-rewriting tool, refining prompts into clearer, structured versions for better image generation.☆3,705Jun 10, 2026Updated last week
- docker-compose-starter☆110Jun 8, 2025Updated last year
- The next generation deep reinforcement learning tookit☆3,463Jun 16, 2023Updated 3 years ago
- A lightweight intelligent agent framework implementing the complete ReAct pattern☆186Aug 15, 2025Updated 10 months ago
- Toolkit for rapid AI and agent prototyping☆149Mar 6, 2026Updated 3 months ago
- [CVPR 2024] Official implementation of "Universal Segmentation at Arbitrary Granularity with Language Instruction"☆283Mar 8, 2024Updated 2 years ago
- Test-Time Augmentation library for Pytorch☆84Jan 2, 2024Updated 2 years ago
- All-in-one third-party firmware for the HoloCubic ESP32 desktop toy. Weather clock, photo/GIF/video player, PC screen mirror, web config,…☆1,024Updated this week
- TVM Documentation in Chinese Simplified / TVM 中文文档☆3,800May 20, 2026Updated 3 weeks ago
- GPU virtual machines on DigitalOcean Gradient AI • AdGet to production fast with high-performance AMD and NVIDIA GPUs you can spin up in seconds. The definition of operational simplicity.
- AI Group is a powerful mobile intelligent assistant application that integrates multiple large language models (LLMs) and AI services, pr…☆1,099Sep 10, 2025Updated 9 months ago
- Open-source framework for automatic video annotation.☆123May 13, 2025Updated last year
- 💰唯一正版💰 minerproxy minerproxy minerproxy minerproxy minerproxy minerproxy minerproxy minerproxy minerproxy minerproxy 矿池抽水 矿池代理 矿池中转 矿池抽…☆3,872Jun 12, 2026Updated last week
- ☆80Mar 5, 2026Updated 3 months ago
- 视频知识管理网站 支持抖音 小红书 b站视频分析管理解析☆163Jun 4, 2026Updated 2 weeks ago
- Run AI models end-to-end encrypted.☆3,155Feb 10, 2025Updated last year
- ☆314Mar 15, 2025Updated last year
- Run frontier LLMs and VLMs with day-0 model support across GPU, NPU, and CPU, with comprehensive runtime coverage for PC (Python/C++), mo…☆8,094Jun 11, 2026Updated last week
- Align Anything: Training All-modality Model with Feedback☆4,661Nov 27, 2025Updated 6 months ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- [CVPR 2026] The official implemention for "PAM: A Pose–Appearance–Motion Engine for Sim-to-Real HOI Video Generation"☆120Mar 21, 2026Updated 2 months ago
- [NeurIPS2025 spotlight★] Official implementation for "RepLDM: Reprogramming Pretrained Latent Diffusion Models for High-Quality, High-Eff…☆298Jan 4, 2026Updated 5 months ago
- SGP41 full-featured driver library for general-purpose MCU and Linux.☆15Oct 26, 2025Updated 7 months ago
- The first open autoregressive foundational video AI model.