Leveraging BERT and c-TF-IDF to create easily interpretable topics.

A bidirectional pipeline parallelism algorithm for computation-communication overlap in DeepSeek V3/R1 training.

Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models

Student version of Assignment 1 for Stanford CS336 - Language Modeling From Scratch

Qwen3Guard is a multilingual guardrail model series developed by the Qwen team at Alibaba Cloud.

[CVPR 2024 Oral] InternVL Family: A Pioneering Open-Source Alternative to GPT-4o. 接近GPT-4o表现的开源多模态对话模型

[NeurIPS 2025] Official website for code and models of Time Series RAG (TS-RAG)

[NeurIPS 2025] Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing

Official source code for AAAI 2025 paper: Augmenting Sequential Recommendation with Balanced Relevance and Diversity

CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

微舆：人人可用的多Agent舆情分析助手，打破信息茧房，还原舆情原貌，预测未来走向，辅助决策！从0实现，不依赖任何框架。

About Code release for "TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis" (ICLR 2023), https://openreview.net/pdf?id=ju_Uqw384Oq

A Library for Advanced Deep Time Series Models for General Time Series Analysis.

📚 从零开始构建大模型

欢迎star⭐。使用微信聊天记录微调大语言模型，使用微信语音消息大模➕0.5B大模型实现高质量声音克隆，并绑定到微信机器人，实现自己的数字分身。 数字克隆/数字分身／声音克隆/LLM/大语言模型/微信聊天机器人/LoRA

A collection of AI tutorials from Dr. Ashish Bamania

Source code of the paper entitled "Exploring Neural Joint Activity in Spiking Neural Networks for Fraud Detection", and presented at CIARP 2024, the 27th Iberamerican Congress on Pattern Recognition.

Dynamic Relation-Attentive Graph Neural Networks for Fraud Detection (ICDMW 2023)

This is a pytorch implementation for the BST model from Alibaba https://arxiv.org/pdf/1905.06874.pdf

A configurable, tunable, and reproducible library for CTR prediction https://fuxictr.github.io

SASRec: Self-Attentive Sequential Recommendation

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

use user APP events to model its behavior and predict

A recommender system model (RecSys) using a transformer neural network. This model takes a sequence of ratings and predicts what a user will rate a new movie. Trained on the movielens dataset.

CIKM 2021: Contrastive Learning of User Behavior Sequence for Context-Aware Document Ranking