740e766e99
refactor: remove summary text from prefix_cache, leave bottom space for subtitles
01ccf192ea
refactor: extract CTA scene into standalone cta.py
c2c18f10d6
fix: replace unicode checkmark with ASCII argmax label
e594d02d31
feat: merge D,E,F into one batch (gated by F at t=8) to show worse static idle
29007c7b97
chore: remove T0-T3 demo tokens from position-grouped decode section
6b0a1dbb5e
refactor: redesign batching FSM as queue pipeline with dynamic task states
c05a432e45
chore: use Times New Roman across all scenes, widen Transformer block to 2.4 for \'Transformer Block × 24\' label
57abefa47f
fix: shift GQA layout down 0.4 to avoid title-input overlap
6b26ec33ab
chore: switch .gitignore to whitelist mechanism (only .py, .md, .gitignore)
0018868ee3
refactor: transformer — heatmap two-phase scores+mask, auto-regressive full I/O pipeline with Emb, RMS Norm, LM Head, distribution
e7d736a3b0
fix: bottom spacing, remove specs card, full formula first with no ellipsis in steps
eeaf0a5a16
feat: SDPA formula breakdown + attention score heatmap with per-cell causal mask
ba100c19d1
refactor: Gantt chart — same 5 requests both sides, wait+run bars, batch boxes connect GPU↔reqs