2025 Technical Notes(1)
Technical notes during 2025 (1).
Summary: Attention is all you need
Summary for paper ‘Attention is all you need’
Summary: A New Golden Age for Computer Architecture
Summary for paper ‘A New Golden Age for Computer Architecture’
2024 Technical Notes
Technical notes during 2024.
GPU Puzzles
This article provides solutions to the GPU Puzzles by Sasha Rush
Tensor Puzzles
This article provides solutions to the Tensor Puzzles by Sasha Rush
Llama_index Source Code Analysis(1)
This blog introduces the basic concepts of RAG and further demonstrates the RAG process based on the source code interpretation of llama_index, including data loader, transformation, index, query, etc. In addition, this paper also analyzes the performance of llama_index RAG process and gives corresponding optimization suggestions.
Llama_index Source Code Analysis(2)
This blog introduces the basic concepts of RAG and further demonstrates the RAG process based on the source code interpretation of llama_index, including data loader, transformation, index, query, etc. In addition, this paper also analyzes the performance of llama_index RAG process and gives corresponding optimization suggestions.
Deep Dive to Pytorch Contiguous Operator(4)
This blog covers PyTorch’s TensorIterator, focusing on fast setup and stride calculation for both normal and ambiguous tensors.
Distributed Training Strategy Introduction
This blog post explores advanced parallelism techniques in deep learning to optimize computational efficiency and memory usage across GPUs. It covers Data Parallelism (DP), Zero Redundancy Optimizer (Zero), Pipeline Parallelism (PP) and Tensor Parallelism (TP).