update

README.md

@@ -11,7 +11,7 @@ thumbnail: https://huggingface.co/spaces/hamoodey/floorplan-graph-rl/resolve/mai
 ---
 **Graded feedback lifts Beam from 0.000 (round 1) to 0.179 held-out partial credit in 10 rounds (strict still 0.000; hand-written reference 0.242): an unsaturated RL environment with room to train.**
 
-**The problem.** Millions of floor plans exist only as images. A room graph (rooms + doors) answers "can a wheelchair reach every room?", "how do I get out?" and "how many rooms?". This environment teaches a text-only model to build that graph from a real plan, and grades it exactly.
+**The problem.** Many floor plans exist only as images or scans. A room graph (rooms + doors) answers "can a wheelchair reach every room?", "how do I get out?" and "how many rooms?". This environment teaches a text-only model to build that graph from a real plan, and grades it exactly.
 
 **In one screen.**
 - Beam is text-only, yet it reads real floor plans by writing a program: 10 rounds of graded feedback took it from **0.000** partial credit in round 1 (practice plans) to **0.179** held-out partial credit on 64 plans it never saw (best round, 9).

index.html

@@ -1,6 +1,6 @@
 <!doctype html><html><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Floorplan Graph RL: Beam reads real floor plans by writing code</title><style>body{font-family:system-ui,-apple-system,sans-serif;max-width:1100px;margin:2em auto;padding:0 1em;line-height:1.5;color:#222}table{border-collapse:collapse;margin:1em 0}td,th{border:1px solid #ccc;padding:6px 10px;text-align:right}td:first-child,th:first-child{text-align:left}.figs{display:grid;grid-template-columns:1fr 1fr;gap:12px}.figs img{width:100%;border:1px solid #ddd}.cap{font-size:.9em;color:#444;margin-bottom:1.5em}code{background:#f3f3f3;padding:1px 4px}.note{color:#666;font-size:.9em}details{margin:.6em 0;border:1px solid #e3e3e3;border-radius:6px;padding:.4em .8em}summary{cursor:pointer;font-size:1.1em}</style></head><body>
 <h1>Floorplan Graph RL: Beam reads real apartment floor plans by writing code</h1>
-<p style='font-size:1.15em'><b>Graded feedback lifts Beam from 0.000 (round 1) to 0.179 held-out partial credit in 10 rounds (strict still 0.000; hand-written reference 0.242): an unsaturated RL environment with room to train.</b></p><p><b>The problem.</b> Millions of floor plans exist only as images. A room graph (rooms + doors) answers "can a wheelchair reach every room?", "how do I get out?" and "how many rooms?". This environment teaches a text-only model to build that graph from a real plan, and grades it exactly.</p><p><b>In one screen.</b></p>
+<p style='font-size:1.15em'><b>Graded feedback lifts Beam from 0.000 (round 1) to 0.179 held-out partial credit in 10 rounds (strict still 0.000; hand-written reference 0.242): an unsaturated RL environment with room to train.</b></p><p><b>The problem.</b> Many floor plans exist only as images or scans. A room graph (rooms + doors) answers "can a wheelchair reach every room?", "how do I get out?" and "how many rooms?". This environment teaches a text-only model to build that graph from a real plan, and grades it exactly.</p><p><b>In one screen.</b></p>
 <ul>
 <li>Beam is text-only, yet it reads real floor plans by writing a program: 10 rounds of graded feedback took it from <b>0.000</b> partial credit in round 1 (practice plans) to <b>0.179</b> held-out partial credit on 64 plans it never saw (best round, 9).</li>
 <li>A 46-line tips playbook <b>more than doubles</b> Beam low: <b>0.075</b> held-out partial credit vs <b>0.032</b> without (3 repeats each; wide spread, one zero in each group).</li>