vlm

【AIGC 实战入门笔记 —— AIGC 摩天大楼】分享 大语言模型（LLMs），大模型高效微调（SFT）,检索增强生成（RAG），智能体（Agent），PPT自动生成, 角色扮演，文生图（Stable Diffusion） ，图像文字识别（OCR），语音识别（ASR），语音合成（TTS），人像分割（SA），多模态（VLM），Ai 换脸(Face Swapping), 文生视频(VD)，图生视频（SVD），Ai 动作迁移，Ai 虚拟试衣，数字人，全模态理解（Omni），Ai音乐生成 干货学习 等 实战与经验。

[ICLR 2025] - Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion

Not just slides - delivery-ready talks

[NeurIPS 2025] Deep Memory Backtracking for Long Video Understanding

vision language models finetuning notebooks & use cases (Medgemma - paligemma - florence .....)

RLLaVA is a user-friendly framework for multi-modal RL research and optimized for resource-constrained teams.

[CVPR 2025] Augmenting Multimodal LLMs with Self-Reflective Tokens for Knowledge-based Visual Question Answering

An AI-powered tool for summarizing YouTube videos by generating scene descriptions, translating them, and creating subtitled videos with text-to-speech narration

Batch Deployment for Document Parsing with AWS Batch & Qwen-2.5-VL

[NeurIPS DB 2025] IR3D-Bench: Evaluating Vision-Language Model Scene Understanding as Agentic Inverse Rendering

Train YOLO + VLM with one command. Auto-generate vision-language training data from YOLO labels - no extra labeling needed.

[CVPR 2026] This is the official PyTorch implementation of "MoDES: Accelerating Mixture-of-Experts Multimodal Large Language Models via Dynamic Expert Skipping".

[ICLR 2026] AgilePruner: An Empirical Study of Attention and Diversity for Adaptive Visual Token Pruning in Large Vision-Language Models

MedMO: Medical Foundation Model

Search knowledge by what documents mean and how they look — not one or the other.

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

Probing the limitations of multimodal language models for chemistry and materials research

Awesome-HCI （Ubiquitous, LLM, MLLM, Agent, RAG, Embodied-AI, RLHF)

[ICLR 2026] - Spectral Concept Selection and Cross-modal Representation Learning for Generalized Category Discovery

Android 16 fork. AI as a platform primitive. Twelve capabilities, one shared runtime, every app. OEM-pluggable. Apache 2.0.