Tag: Performance Optimization
-
Pillow to OpenCV: 3.4x Faster Image Load in Production
Pillow's imread took 800ms per image. OpenCV cut it to 240ms. Here's the color space trap, dtype bug, and EXIF hell that break every migration.
-
Adjacency List vs Matrix: 47% Faster DFS in Interview Code
DFS performance comparison: adjacency list beats matrix by 47% in interview problems. See which graph representation wins and why it matters.
-
PaddleOCR Optimization: Caching, Warmup, and Quantization Benchmarks
PaddleOCR ocr.predict vs ocr.ocr performance compared. Benchmark caching, warmup, and INT8 quantization โ init time drops 73% with the right setup.