TildAlice – Tech Blog for Developers
Python, AI, Deep Learning, Data Analysis, Computer Vision, Reinforcement Learning โ tips and tutorials
Latest Posts
-
RLHF๋ก ์ธ์ด๋ชจ๋ธ ์ ๋ ฌํ๊ธฐ: ChatGPT๋ถํฐ Claude๊น์ง์ ์ค์ ๊ตฌํ ๊ฐ์ด๋
RLHF implementation guide: train ChatGPT-style models with human feedback. PPO + reward model code walkthrough included.
-
Ruff์ uv๋ก Python ํ๋ก์ ํธ ๋น๋ ์๊ฐ 90% ๋จ์ถํ๊ธฐ: Rust ๊ธฐ๋ฐ ํด์ฒด์ธ ์์ ์ ๋ณต
Ruff + uv vs traditional tools: 90% faster linting and 10x faster installs. Benchmark results from migrating 50K line codebase.
-
๋ธ๋ก๊ทธ ์์
DevTips tech blog launch: documenting coding challenges, solutions, and growth journey. First post on why developers should blog.
-
ViT vs CNN Attention Map ๋น๊ต: ๋ชจ๋ธ ํด์ 5๊ฐ์ง ๊ธฐ๋ฒ
ViT vs CNN attention maps decoded: 5 techniques to visualize what your model actually sees. GradCAM, rollout, and flow compared.
-
Polars vs Pandas: ๋์ฉ๋ ๋ฐ์ดํฐ ์ฒ๋ฆฌ ์ฑ๋ฅ ๋น๊ต์ ๋ง์ด๊ทธ๋ ์ด์ ๊ฐ์ด๋
Polars beats Pandas by 10-100x on large datasets. Benchmark results + migration guide for switching without breaking your pipeline.
-
RLHF vs DPO vs KTO: LLM ์ ๋ ฌ(Alignment) ๊ธฐ๋ฒ ์๋ฒฝ ๋น๊ต ๊ฐ์ด๋
RLHF, DPO, or KTO for LLM alignment? Compare training costs, data needs, and performance on real tasks to pick the right method.
-
ViT vs ConvNeXt: 2026๋ ์ด๋ฏธ์ง ๋ถ๋ฅ ๋ชจ๋ธ ์ํคํ ์ฒ ์ ํ ๊ฐ์ด๋
ViT or ConvNeXt for image classification in 2026? Speed vs accuracy tradeoffs tested on ImageNet with deployment recommendations.
-
Python 3.13 Per-Interpreter GIL๋ก ์ง์ง ๋ณ๋ ฌ ์ฒ๋ฆฌ ๊ตฌํํ๊ธฐ – ์ฑ๋ฅ 2๋ฐฐ ํฅ์ ์ค์ ๊ฐ์ด๋
Python 3.13 Per-Interpreter GIL doubles CPU performance for parallel tasks. Benchmarks + code showing how to escape the GIL forever.