At TiPaiPai, we focus on building engineering teams and culture, cultivating technical insights and practice, and fostering sharing, growth, and connection.
This article explains how Go executables receive and process command‑line arguments and environment variables, compares C and Go implementations, dives into the Linux stack layout, and details the flag package's parsing mechanics with code examples and runtime internals.
This article explains how attention mechanisms are applied to text recognition, covering the basic CNN‑Seq2Seq‑Attention architecture, multi‑scale attention extensions, and a 2D attentional irregular scene text recognizer with detailed network components, training loss, and experimental results.
This tutorial walks through the evolution from classic Attention mechanisms to Transformers, presents detailed code implementations, and demonstrates how to integrate a Transformer unit into a text recognition network for enhanced visual and semantic feature extraction.
The article introduces the Vue v‑timely‑validate directive, explains how it shifts form validation from blur to change for faster error clearance, outlines form structure, validation principles, usage steps, and provides complete code examples for seamless integration into i‑form components.
This article reviews a CVPR 2018 paper that introduces a dense‑connected encoder and multi‑scale attention mechanism to improve handwritten mathematical expression recognition, detailing the background, network architecture, GRU decoder, loss function, and experimental gains over previous methods.
This article reviews two recent computer‑vision approaches—DRRG and STKM—that combine CNN backbones with graph‑based relational reasoning and self‑attention to achieve state‑of‑the‑art detection of arbitrarily shaped text in images.
This article explains the challenges of scene text detection and introduces two state‑of‑the‑art models, ContourNet and CenterNet, detailing their architectural innovations, loss functions, and how they overcome issues like extreme aspect ratios and anchor‑based inefficiencies.
This article reviews two end‑to‑end deep‑learning approaches—DocUNet (CVPR 2018) and DewarpNet (ICCV 2019)—for correcting warped document images, detailing their network architectures, synthetic data generation, loss functions, experimental results, and the remaining challenges in document dewarping.
In this reflective interview, a senior front‑end specialist shares how breaking technical silos, continuous learning, disciplined work habits, mentorship, and personal resilience shape his engineering mindset and career development, offering practical wisdom for developers seeking lasting growth.
This article introduces TensorRT, explains its deployment workflow from model training to engine generation, shows how to register custom operators for ONNX and create TensorRT plugins, and explores deformable convolution (DCN) implementation strategies for high‑performance AI inference.
