TildAlice – Tech Blog for Developers
Python, AI, Deep Learning, Data Analysis, Computer Vision, Reinforcement Learning โ tips and tutorials
Latest Posts
-
LoRA vs QLoRA vs Full Fine-tuning: GPU Memory Benchmarks
Full fine-tuning costs $5/hr on A100. QLoRA drops it to $0.50 on T4 โ with matching accuracy at rank 64. Real memory breakdowns + 47-run benchmark.
-
tmux vs zellij: 100-Pane Split Render Speed Benchmark
Benchmark tmux vs zellij rendering 100 split panes. Which terminal multiplexer wins on speed? Data-driven comparison reveals the clear winner.
-
VS Code vs Neovim: Python Startup Speed & LSP Benchmarks
Compare Python LSP startup times in VS Code vs Neovim. Benchmark results reveal which editor wins for speedโplus tips to optimize both setups.
-
GPT-4 vs Claude 3.5 vs Gemini: MMLU Zero-Shot Accuracy
GPT-4 beats Claude 3.5 by just 1.8% on zero-shot MMLU โ way closer than official benchmarks claim. Real accuracy numbers from 1,000 questions.
-
import * in Functions: UnboundLocalError & Namespace Fix
Learn why `import *` inside functions causes UnboundLocalError and discover the namespace mechanics behind this Python gotcha with practical fixes.
-
gdb + py-spy vs pdb: Production Debug Latency Compared
gdb and py-spy attach to live Python processes without restarts. Compare overhead, introspection depth, and when sampling beats breakpoints.
-
Kubernetes Autoscaler for ML: Build HPA from Scratch
Build Kubernetes HPA for ML workloads from scratch. Learn custom metrics, scaling strategies, and real-world GPU optimization patterns.
-
GNN for Multi-Component RUL: Graph Attention Beats Fixed Topology
Learn how Graph Attention Networks predict multi-component system failures better than fixed-topology GNNs. Attention mechanism key to RUL accuracy.