深度学习基础
深度学习:维度灾难(Curse Of Dimensionality)
机器学习
机器学习:概率视角的线性回归(Linear Regression)
机器学习:随机梯度下降(SGD)与梯度下降(GD)的区别与代码实现。
深度学习:手写反向传播算法(BackPropagation)与代码实现
计算机视觉
深度学习:NiN(Network In Network)详细讲解与代码实现
深度学习: Batch Normalization论文详细解读
目标检测:RCNN、SppNet、Fast RCNN、Faster RCNN是如何过渡的?
目标检测:特征金字塔网络(Feature Pyramid Network)
自动驾驶
BEV感知:BEV开山之作LSS(lift,splat,shoot)原理代码串讲
多模态
CLIP
多模态速读:ViLT、ALBEF、VLMO、BLIP
LLM
深度学习:Self-Attention与Multi-heads Attention详解
ChatGLM基座:GLM(General Language Model)论文阅读笔记
DeepNet :Scaling Transformers to 1000 Layer
Roformer:Enhanced Transformer with rotary position embedding
Lora:Low-Rank Adapation of Large Language models
LLaMA:Open and Efficient Foundation Language Models
Self-Instruct:Aligning Language Model with Self Generated Instructions
Prefix-Tuning: Optimizing Continuous Prompts for Generation
P-Tuning : GPT Understands,Too 论文笔记
P-Tuning v2: Prompt Tuning Can Be Comparable to Fine-tuning Universally Across Scales and Tasks
InstructGPT:Training language models to follow instrcutions with human feedback
huggingface TRL是如何实现20B-LLM+Lora+RLHF
LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-into Attention论文解读
Prefix-tuning、Adapter、LLaMA-Adapter的流程图与伪代码实现
技术报告:Efficient and Effective Text Encoding for Chinese LLaMA AND Alpaca
Huggingface的GenerationConfig 中的top_k与top_p详细解读
基于人类反馈的强化学习(RLHF)在LLM领域是如何运作的?
WizardKM:Empowering Large Language Models to Follow Complex Instructions
Distilling Step-by-Step: 可以用更少的训练数据与模型尺寸战胜同级别的LLM!
加速训练
Zero系列三部曲:Zero、Zero-Offload、Zero-Infinity
大数据
大数据基础:HDFS(分布式文件系统)前置知识,吞吐量,数据块,并发
DEBUG
Pytorch:关于nn.dataParallel我所踩过的坑
Python 程序设计
Python:Math.inf是什么意思?
文章评论