LLM Course: LLM Curriculum with Roadmaps and Colab Notebooks 182

A curated collection of LLM courses with learning roadmaps and Colab notebooks, helping users build knowledge and skills in LLMs from foundational to advanced levels.

#Ai工具箱 #Ai学习资源

Visit LLM Course: LLM Curriculum with Roadmaps and Colab Notebooks

LLM Course: LLM Curriculum with Roadmaps and Colab NotebooksIntroduction

What is LLM Course?

LLM Course provides a collection of courses on LLMs, including learning roadmaps and Colab notebooks to help users gradually master the knowledge and applications of LLMs from basic to advanced levels. The courses cover mathematics, Python programming, neural networks, natural language processing (NLP), LLM architecture, pretraining, fine-tuning, preference alignment, evaluation, quantization, and deployment.

LLM Course: A collection of LLM courses including roadmaps and Colab notebooks.webp

Main Content of LLM Course

LLM Fundamentals

Mathematics Fundamentals: Linear algebra, calculus, probability, and statistics.
Python Programming: Python basics, data science libraries, data processing, machine learning libraries.
Neural Networks: Fundamentals of neural networks, training and optimization, overfitting and solutions.
Natural Language Processing (NLP): Text preprocessing, feature extraction techniques, word embeddings, recurrent neural networks (RNNs).

LLM Scientist

LLM Architecture: Transformer architecture, attention mechanisms, sampling strategies.
Pre-training Models: Data preparation, distributed training, training optimization, monitoring.
Post-training Datasets: Data structures, storage and chat templates, synthetic data generation, data augmentation, quality filtering.
Supervised Fine-Tuning (SFT): Training techniques, parameter settings, distributed training, monitoring.
Preference Alignment: Direct Preference Optimization (DPO), Proximal Policy Optimization (PPO), monitoring.
Evaluation: Automated benchmarking, human evaluation, model evaluation, feedback signals.
Quantization: Basic techniques, GGUF & llama.cpp, GPTQ & AWQ, SmoothQuant & ZeroQuant.
New Trends: Model merging, multimodal models, interpretability, test-time compute.

LLM Engineer

Running LLMs: LLM APIs, open-source LLMs, prompt engineering, structured output.
Building Vector Stores: Document loading, splitting, embedding models, vector databases.
Retrieval-Augmented Generation (RAG): Orchestrators, retrievers, memory, evaluation.
Advanced RAG: Query construction, agents and tools, post-processing, programmatic LLMs.
Inference Optimization: Flash Attention, key-value caching, speculative decoding.
Deploying LLMs: Local deployment, demo deployment, server deployment, edge deployment.
Securing LLMs: Prompt injection, backdoor attacks, defense measures.

LLM Course.webp

Tools and Resources

Tools: LLM AutoEval, LazyMergekit, LazyAxolotl, AutoQuant, Model Family Tree, ZeroSpace.
Resources: 3Blue1Brown videos, StatQuest tutorials, Khan Academy courses, Real Python tutorials, freeCodeCamp videos, and more.