Spaces:
Running
Running
Download index.html from thelamapi/Next-Signal: direct link, hf CLI and curl.
- Browser
- Download file 3.13 kB
-
https://huggingface.co/spaces/thelamapi/Next-Signal/resolve/main/index.html
- Command line
-
hf download hf://spaces/thelamapi/Next-Signal/index.html
-
curl -L -o index.html https://huggingface.co/spaces/thelamapi/Next-Signal/resolve/main/index.html
3.13 kB
| <html lang="en"> | |
| <head> | |
| <meta charset="utf-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1"> | |
| <link rel="icon" href="data:,"> | |
| <title>Next Signal</title> | |
| <link rel="stylesheet" href="style.css"> | |
| </head> | |
| <body> | |
| <main> | |
| <header> | |
| <h1>Next Signal</h1> | |
| <p>An encoder learned from <em>unlabeled</em> smartphone sensor windows, tested with a frozen linear probe. Pick a window from people the model never saw, distort it, or upload your own; three representations classify it in your browser.</p> | |
| <div class="badges"><span class="badge">Live in-browser inference</span><span class="badge">UCI HAR (CC BY 4.0)</span><span class="badge">Test subjects held out</span><a class="badge" href="https://huggingface.co/thelamapi/Next-Signal">Model repo</a></div> | |
| </header> | |
| <div class="grid"> | |
| <section class="card"> | |
| <h2>Input window (2.56 s, 9 channels)</h2> | |
| <label for="win">Held-out test window</label><select id="win"></select> | |
| <label for="file">…or upload a CSV: 128 rows × 9 columns in the UCI HAR channel order</label><input type="file" id="file" accept=".csv,.txt"> | |
| <label for="noise">Added noise (× channel std): <b id="noise_v"></b></label><input type="range" id="noise" min="0" max="1.5" step="0.1" value="0"> | |
| <label for="scale">Amplitude scale: <b id="scale_v"></b></label><input type="range" id="scale" min="0.5" max="1.5" step="0.05" value="1"> | |
| <label for="shift">Time shift (samples): <b id="shift_v"></b></label><input type="range" id="shift" min="0" max="127" step="1" value="0"> | |
| <label for="budget">Probe trained with labels per class</label> | |
| <select id="budget"><option value="2">2</option><option value="10" selected>10</option><option value="50">50</option></select> | |
| <p class="status" id="status"></p> | |
| </section> | |
| <section class="card"> | |
| <h2>Signal</h2> | |
| <select id="chans" style="max-width:220px"><option value="acc">body acceleration</option><option value="gyro">gyroscope</option><option value="total">total acceleration</option></select> | |
| <canvas id="sig" style="margin-top:6px"></canvas><div id="sigl"></div> | |
| <p class="small muted" id="truth"></p> | |
| </section> | |
| </div> | |
| <div class="grid" style="margin-top:12px"> | |
| <section class="card"><h2>Self-supervised encoder + probe</h2><div id="p_ssl"></div></section> | |
| <section class="card"><h2>Random encoder + probe</h2><div id="p_rand"></div></section> | |
| <section class="card"><h2>Hand-made statistics + probe</h2><div id="p_stats"></div></section> | |
| </div> | |
| <p class="small muted">Inference took <span id="ms">–</span> ms for all three. Every probe saw the same labeled windows.</p> | |
| <section class="card" style="margin-top:12px"><h2>Recorded benchmark</h2><img src="preview.png" alt="Probe accuracy vs labels" style="width:100%;border-radius:8px"></section> | |
| <footer>Research demo on public smartphone data from 30 volunteers; not a medical or safety device.</footer> | |
| </main> | |
| <script src="nlrt.js"></script> | |
| <script src="nlui.js"></script> | |
| <script src="core.js"></script> | |
| <script src="app.js"></script> | |
| </body> | |
| </html> | |