85 min read
Vision Transformer Deep Dive: ViT to Modern Descendants
Build ViT from pixels, then trace DeiT, Swin, DINOv2, SAM, LLaVA, and practical latency optimizations.
Deep Learning
Explore technical articles related to Deep Learning. Find in-depth analysis, tutorials, and insights.
- Tagged with
- 11 articles
Articles Related to Deep Learning
25 min read
CUDA Matrix Multiplication Optimization: From Naive to Near-cuBLAS
Step-by-step CUDA matrix multiplication optimization with 9 interactive visualizations. From naive kernels through shared memory tiling to near-cuBLAS speeds.
25 min read
Numerical Sensitivity: Why FP16 Breaks NAdam
Visual exploration of floating-point arithmetic and numerical stability. Learn why NAdam fails in FP16 and how machine epsilon affects deep learning.
12 min read
SAM's Multi-Mask Ambiguity: A Visual Deep Dive
Deep dive into how SAM resolves point prompt ambiguity through three-mask output design, IoU prediction, and intelligent mode switching.
15 min read
YOLOv11 Loss Functions Explained: Interactive Visual Guide
Understand YOLOv11's loss functions through interactive visualizations. Compare IoU variants (GIoU, DIoU, CIoU), explore Distribution Focal Loss (DFL), and see why anchor-free detection matters.
16 min read
How torch.compile Generates Optimized GPU Kernels: Fusion, Tiling, and Shape Specialization
How torch.compile generates optimized GPU kernels from PyTorch eager code: kernel fusion, memory tiling, shape specialization, and the TorchInductor backend — with interactive visualizations of every transformation.
7 min read
Fix PyTorch "view size is not compatible" Error
Learn why PyTorch throws the "view size is not compatible" error, understand tensor memory layout, and discover optimal solutions with performance benchmarks.
15 min read
GGML File Structure: Tensor Layout and Quantized Model Format Explained
Understand GGML file structure and quantization formats used by local LLMs. Visual guide to how llama.cpp stores and loads model weights efficiently.
30 min read
Quantization Deep Dive: From FP32 to INT4
Master neural network quantization with interactive visualizations. Explore QAT, PTQ, GPTQ, AWQ, and SmoothQuant methods for efficient model deployment.
25 min read
How TensorRT Works: NVIDIA Inference Optimization
Explore TensorRT optimization: layer fusion, INT8 quantization, kernel auto-tuning, and deployment strategies with 8+ interactive visualizations.
15 min read
Kernel Fusion in Deep Learning: How GPU Kernels Are Merged
Kernel fusion merges multiple neural network operations into a single GPU kernel to eliminate intermediate memory writes — this article explains how fusion works, why it helps deep learning workloads, and how TensorRT and torch.compile use it.