GuanYuXx/DQN-variants-demo

HW3-3: Randomize button now reshuffles the Validation grid

UI: remove hero cards, relocate Randomize button to bottom action bar

P1 + SITE_OPTIMIZATION Phase 1+2: sync_freq, sticky nav, hero, KaTeX, arch diagrams

Compare mode UI: relocate action bar, add Random button + Player drag

HW3-1: add Target Network to Basic mode + mark bonus items in UI/README

Untrack __pycache__ build artifacts

Add Rainbow DQN mode (Double + Dueling + PER + Multi-step + NoisyNet + C51)

Compare mode: switch to offline-pretrained Double/Dueling DQN

Add HF Spaces README config

Track binary files with Git LFS

Add deployment config for Hugging Face Spaces

版面調整

Add requirements.txt for environment setup

Implement offline training with PER, Double DQN, and multi-obstacle Gridworld support

Update: Change chart update frequency back to 1 epoch for smoother visuals

Update: Add drag-and-drop support, randomize button, and 10-epoch chart update frequency

Update: Add real-time SSE loss tracking and separate layout for buffer route drawing and final animation

Initial commit: DQN with Experience Replay, Flask Backend, and Frontend