OpenVLA: An open-source vision-language-action model for robotic manipulation.

Recent Advances on MLLM's Reasoning Ability

A framework for few-shot evaluation of language models.

[NeurIPS '18] "Can We Gain More from Orthogonality Regularizations in Training Deep CNNs?" Official Implementation.

Preventing Dimensional Collapse in Self-Supervised Learning via Orthogonality Regularization

Search-o1: Agentic Search-Enhanced Large Reasoning Models

Source code for models described in the paper "AudioCLIP: Extending CLIP to Image, Text and Audio" (https://arxiv.org/abs/2106.13043)

[CVPR 2025] LamRA: Large Multimodal Model as Your Advanced Retrieval Assistant

🚀 Efficient implementations of state-of-the-art linear attention models in Torch and Triton

[AAAI Alignment Track 25 Poster] Stream Aligner: Efficient Sentence-Level Alignment via Distribution Induction

小红书接口，爬虫用

Experiments on Multi-Head Latent Attention

[CVPR 2024] Official PyTorch implementation of FreeCustom: Tuning-Free Customized Image Generation for Multi-Concept Composition

ELITE: Encoding Visual Concepts into Textual Embeddings for Customized Text-to-Image Generation (ICCV 2023, Oral)

EGSRAL: An Enhanced 3D Gaussian Splatting based Renderer with Automated Labeling for Large-Scale Driving Scene

[ICASSP 2025] This is the official repository of Semantic Hierarchical Prompt Tuning for Parameter-Efficient Fine-Tuning

Chinese safety prompts for evaluating and improving the safety of LLMs. 中文安全prompts，用于评估和提升大模型的安全性。

NeurIPS Large Language Model Efficiency Challenge: 1 LLM + 1GPU + 1Day

Train a 1B LLM with 1T tokens from scratch by personal

A Survey on LLM Agents (Published on CoLing 2025)

Collection of training data management explorations for large language models

This is official implementation of FedTP (TNNLS.2023) [https://ieeexplore.ieee.org/abstract/document/10130784].

"LightRAG: Simple and Fast Retrieval-Augmented Generation"

The codes and datasets about our ACL 2024 Main Conference paper titled "Cognitive Visual-Language Mapper: Advancing Multimodal Comprehension with Enhanced Visual Knowledge Alignment"