TildAlice – Tech Blog for Developers
Python, AI, Deep Learning, Data Analysis, Computer Vision, Reinforcement Learning โ tips and tutorials
Latest Posts
-
Recursive DFS vs Iterative Stack: Cycle Detection Performance and Limits
Recursive DFS is 30% faster but crashes at 10K nodes. Iterative survives 50K graphs โ here's the stack overhead breakdown and when each wins.
-
Stable Baselines3 VecEnv Reset Bug: 100K Step Desync Fix
Fix the Stable Baselines3 VecEnv reset bug causing 100K step desyncs. Learn why auto_reset breaks training and how to solve it properly.
-
RLHF in 2026: Why Human Feedback Still Beats Pure AI Alignment
Explore RLHF in 2026 and discover why human feedback remains essential for AI alignmentโeven as models grow more capable. The surprising reasons inside.
-
SAM 2 Inference Pipeline Bottlenecks: 3x Slower Than SAM
SAM 2 inference runs 3x slower than SAM despite better accuracy. Discover the bottleneck causing this performance drop and how to optimize it.
-
OpenCV Face Blur: 3 Mistakes That Kill Real-Time FPS
Haar cascade + Gaussian blur drops your FPS to 15? Here's the numpy slice bug, kernel size trap, and detection parameter fix that gets you back to 30.
-
uv vs pip vs Poetry: Install Speed Test on 50+ Packages
pip took 68s, Poetry 227s, uv just 4s for 52 packages. Here's why the Rust-based installer wins for beginners โ and when to stick with Poetry.
-
Mamba-2 vs Mamba vs Transformer: Long Range Arena Results
Mamba-2 claims 8x faster training than Mamba while matching accuracy on 16K-token tasks. Here's what the Long Range Arena benchmark reveals.
-
Why I Switched from Ledger to Open Source Hardware Wallets
Ledger's closed firmware and Recover controversy made me rethink security. Here's why open source hardware wallets like OneKey matter for self-custody.