Spaces:
Running on Zero
Running on Zero
| """HORST: Deutsche Rechtschreib- und Grammatikkorrektur mit einem eigenen Modell (ZeroGPU).""" | |
| import base64 | |
| from pathlib import Path | |
| import gradio as gr | |
| import sentencepiece as spm | |
| import spaces | |
| import torch | |
| from huggingface_hub import hf_hub_download | |
| from evaluate import correct_text, load_model | |
| from prepare import detokenize, split_sentences, tokenize | |
| from ui_helpers import EMPTY_DIFF, MAX_WORDS, counter_html, diff_html, stats_html | |
| HERE = Path(__file__).parent | |
| LOGO = "data:image/svg+xml;base64," + base64.b64encode((HERE / "horst-wordmark.svg").read_bytes()).decode() | |
| REPO = "Bayernator/horst-ai" | |
| # ZeroGPU emuliert CUDA außerhalb von @spaces.GPU, das Modell gehört laut Doku auf Modulebene auf "cuda" | |
| device = "cuda" if torch.cuda.is_available() else "cpu" | |
| model = load_model([hf_hub_download(REPO, "horst-ai.pt")], device) | |
| sp = spm.SentencePieceProcessor(model_file=hf_hub_download(REPO, "spm.model")) | |
| def duration(text, beam): | |
| return min(120, 15 + int(len((text or "").split()) * 0.2)) # kurze Dauer = bessere Warteschlangen-Priorität | |
| def correct(text, beam): | |
| text = (text or "").strip() | |
| if not text: | |
| return "" | |
| if len(text.split()) > MAX_WORDS: | |
| raise gr.Error(f"Bitte höchstens {MAX_WORDS} Wörter auf einmal eingeben ({len(text.split())} sind zu viel).") | |
| out = [] | |
| for line in text.split("\n"): # Absätze bleiben erhalten | |
| sents = [tokenize(s) for s in split_sentences(line) if s.strip()] | |
| out.append(detokenize(correct_text(model, sp, sents, beam=int(beam))) if sents else "") | |
| return "\n".join(out) | |
| def show_diff(old, new): | |
| if not (new or "").strip(): | |
| return EMPTY_DIFF, result_head() | |
| html, n = diff_html(old, new) | |
| return html, result_head(stats_html(new, n)) | |
| def result_head(chips=""): | |
| return f'<div class="px-row"><span class="px-kicker">Ergebnis</span>{chips}</div>' | |
| def clear(): | |
| return "", "", EMPTY_DIFF, result_head(), counter_html("") | |
| ARROW = ('<svg width="18" height="18" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.4" ' | |
| 'stroke-linecap="square" aria-hidden="true"><path d="M12 5v14M6 13l6 6 6-6"/></svg>') | |
| def pixels(*colors): | |
| return "".join(f'<i style="background:{c}"></i>' for c in colors) | |
| MARK = pixels("#FFD23F", "#FF8A1F", "#F2601C", "#D9381E") | |
| HERO = f""" | |
| <header class="px-nav"> | |
| <div class="px-brand"><span class="px-mark" aria-hidden="true">{MARK}</span>HORST</div> | |
| <nav class="px-links" aria-label="Links"> | |
| <a href="https://huggingface.co/Bayernator/horst-ai" target="_blank">Modell</a> | |
| <a href="https://bavarianator.github.io/horst/benchmark.html" target="_blank">Benchmark</a> | |
| <a href="https://bavarianator.github.io/horst/" target="_blank">Website</a> | |
| <a class="px-btn dark" href="#korrektur">Ausprobieren</a> | |
| </nav> | |
| </header> | |
| <section class="px-hero"> | |
| <div> | |
| <span class="px-new">Neu: Trainingsstufe 2</span> | |
| <h1 class="px-h1">Deutsche Text­korrektur. Von Grund auf <mark>selbst trainiert.</mark></h1> | |
| <p class="px-lead">Rechtschreibung, Groß- und Kleinschreibung, Kommas und Grammatik: korrigiert von einem KI-Modell, | |
| das komplett von Grund auf selbst trainiert wurde.</p> | |
| <div class="px-cta"> | |
| <a class="px-btn dark" href="#korrektur">Text korrigieren {ARROW}</a> | |
| <a class="px-btn line" href="https://huggingface.co/Bayernator/horst-ai" target="_blank">Modell ansehen</a> | |
| </div> | |
| <div class="px-facts"> | |
| <div class="px-fact"><b>209 Mio.</b><span>Parameter</span></div> | |
| <div class="px-fact"><b>0</b><span>fremde Gewichte</span></div> | |
| <div class="px-fact"><b>{MAX_WORDS}</b><span>Wörter am Stück</span></div> | |
| <div class="px-fact"><b>GPU</b><span>ZeroGPU</span></div> | |
| </div> | |
| </div> | |
| <div class="px-poster"> | |
| <span class="px-kicker">Deutsche Rechtschreib- und Grammatikkorrektur</span> | |
| <img src="{LOGO}" alt="HORST"> | |
| <div class="px-stair small" aria-hidden="true"></div> | |
| </div> | |
| </section> | |
| <div class="px-stair" aria-hidden="true"></div> | |
| <div class="px-head"> | |
| <div><span class="px-kicker">01 / Korrektur</span><h2 class="px-h2">Text rein. Fehler raus.</h2></div> | |
| <span class="px-count">Bis {MAX_WORDS} Wörter · Absätze bleiben erhalten</span> | |
| </div> | |
| """ | |
| def pbar(label, value, stage2=False): | |
| n = round(value * 20) # 20 Pixel = F0,5 von 1,0 | |
| shown = f"{value:.2f}".replace(".", ",") | |
| return (f'<div class="px-bar"><span>{label}</span><div class="px-pbar{" s2" if stage2 else ""}" style="--n:{n}"></div>' | |
| f'<span>{shown}</span></div>') | |
| def metric(big, label, src, s1, s2): | |
| return (f'<div class="px-metric"><div class="px-big">{big}</div><p class="px-mlabel">{label}</p><p class="px-msrc">{src}</p>' | |
| f'<div class="px-bars">{pbar("Stufe 1", s1)}{pbar("Stufe 2", s2, True)}</div></div>') | |
| INFO = f""" | |
| <section class="px-model"> | |
| <span class="px-kicker">02 / Neu im Space</span> | |
| <h2 class="px-h2">Trainingsstufe 2 ist da.</h2> | |
| <p class="px-sub">Gemischte Daten mit echten Korrekturen aus Wikipedia und Lernertexten, dazu lange Passagen bis 500 Wörter. | |
| Gleiche Modellgröße, deutlich bessere Korrekturen.</p> | |
| <div class="px-metrics"> | |
| {metric("0,63", "Echte Lernertexte", "Falko-MERLIN · dev", 0.40, 0.633)} | |
| {metric("0,88", "Harte Rechtschreibfälle", "german-ortho · bench-dev", 0.69, 0.879)} | |
| <div class="px-metric"><span class="px-kicker">Steckbrief</span><div class="px-specs"> | |
| <div class="px-spec"><span>Parameter</span><b>209 Mio.</b></div> | |
| <div class="px-spec"><span>Kontext</span><b>1024 Tokens</b></div> | |
| <div class="px-spec"><span>Stufe 1</span><b>256 Tokens</b></div> | |
| <div class="px-spec"><span>Trainingsschritt</span><b>32.000</b></div> | |
| </div></div> | |
| </div> | |
| <p class="px-cap">F0,5 auf Edit-Ebene, je 200 Beispiele, greedy (Beam 1). Höher ist besser.</p> | |
| </section> | |
| <section class="px-section"> | |
| <span class="px-kicker">03 / So entsteht HORST</span> | |
| <div class="px-steps"> | |
| <div class="px-step"><div class="px-num" aria-hidden="true">{pixels("#FFD23F")}</div><h3>Von Grund auf trainiert</h3> | |
| <p>Keine fremden Gewichte: Tokenizer und Transformer sind komplett selbst gebaut und trainiert.</p></div> | |
| <div class="px-step"><div class="px-num" aria-hidden="true">{pixels("#FFD23F", "#FF8A1F")}</div> | |
| <h3>Millionen Übungssätze</h3> | |
| <p>30 Mio. saubere deutsche Sätze mit künstlichen Fehlern. In Stufe 2 kommen echte Korrekturen aus Wikipedia-Versionen | |
| und Lernertexten dazu.</p></div> | |
| <div class="px-step"><div class="px-num" aria-hidden="true">{pixels("#FFD23F", "#FF8A1F", "#D9381E")}</div><h3>Auch für lange Texte</h3> | |
| <p>Absätze bleiben erhalten. Stufe 2 liest bis zu 1024 Tokens am Stück und korrigiert mit Kontext aus den Nachbarsätzen.</p></div> | |
| </div> | |
| <div class="px-note"> | |
| <svg width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="square" | |
| aria-hidden="true"><rect x="3" y="3" width="18" height="18"/><path d="M12 11v6M12 7v1"/></svg> | |
| <span>HORST ist ein Forschungsmodell (Trainingsstufe 2) und macht noch Fehler. Prüfe wichtige Texte immer selbst.</span> | |
| </div> | |
| </section> | |
| <footer class="px-foot"> | |
| <div class="px-stair" aria-hidden="true"></div> | |
| <div class="px-foot-in"> | |
| <img src="{LOGO}" alt="HORST"> | |
| <nav aria-label="Weiterführend"> | |
| <a href="https://huggingface.co/Bayernator/horst-ai" target="_blank">Modell</a> | |
| <a href="https://huggingface.co/datasets/Bayernator/german-spelling-data" target="_blank">Trainingsdaten</a> | |
| <a href="https://huggingface.co/datasets/Bayernator/german-ortho" target="_blank">Benchmark-Daten</a> | |
| <a href="https://bavarianator.github.io/horst/" target="_blank">Website</a> | |
| </nav> | |
| </div> | |
| </footer> | |
| """ | |
| # Pixeltreppe: Zeilen per Hintergrund (von rechts), Pixelspalten per Maske | |
| STAIR = ("linear-gradient(#FFD23F,#FFD23F) right 0/{0} 18px no-repeat,linear-gradient(#FFB020,#FFB020) right 20px/{1} 18px no-repeat," | |
| "linear-gradient(#FF8A1F,#FF8A1F) right 40px/{2} 18px no-repeat,linear-gradient(#F2601C,#F2601C) right 60px/{3} 18px no-repeat," | |
| "linear-gradient(#D9381E,#D9381E) right 80px/{4} 18px no-repeat") | |
| PIXELS = ("-webkit-mask-image:linear-gradient(90deg,#000 0 {0}px,transparent {0}px);-webkit-mask-size:{1}px 100%;-webkit-mask-repeat:repeat-x;" | |
| "mask-image:linear-gradient(90deg,#000 0 {0}px,transparent {0}px);mask-size:{1}px 100%;mask-repeat:repeat-x") | |
| CSS = """ | |
| .gradio-container { max-width: 1200px !important; margin: 0 auto !important; padding: 0 32px !important; } | |
| .px-kicker, [data-testid="block-info"] { font-family: "IBM Plex Mono", ui-monospace, monospace !important; font-size: 12.5px !important; | |
| font-weight: 500 !important; letter-spacing: .12em; text-transform: uppercase; color: #4B453E !important; } | |
| .px-stair { height: 118px; -webkit-mask-position: right top; mask-position: right top; | |
| background: STAIR_BIG, linear-gradient(#161412,#161412) right 100px/100% 18px no-repeat; PIXELS_BIG; } | |
| .px-stair.small { height: 98px; margin: 0 -36px; background: STAIR_SMALL; } | |
| .px-foot .px-stair { height: 98px; background: STAIR_FOOT; } | |
| .px-nav { display: flex; align-items: center; justify-content: space-between; gap: 20px; flex-wrap: wrap; padding: 30px 0 22px; } | |
| .px-brand { display: flex; align-items: center; gap: 10px; font-weight: 800; font-size: 22px; letter-spacing: .02em; } | |
| .px-mark { display: grid; grid-template-columns: repeat(2, 10px); gap: 2px; } | |
| .px-mark i, .px-num i { display: block; width: 10px; height: 10px; } | |
| .px-links { display: flex; align-items: center; gap: 26px; flex-wrap: wrap; font-size: 15px; font-weight: 500; } | |
| .px-links a { color: #161412; text-decoration: none; } | |
| .px-links a:hover { color: #C02E15; } | |
| .px-btn { display: inline-flex; align-items: center; justify-content: center; gap: 10px; min-height: 48px; padding: 0 22px; | |
| border: 1.5px solid #161412; font-weight: 700; font-size: 16px; text-decoration: none !important; } | |
| .px-btn.dark, .px-links a.px-btn { background: #161412; color: #FBF6EC !important; } | |
| .px-btn.dark:hover { background: #F2601C; border-color: #F2601C; color: #161412 !important; } | |
| .px-btn.line { color: #161412; } | |
| .px-btn.line:hover { background: #161412; color: #FBF6EC; } | |
| .px-hero { display: grid; grid-template-columns: minmax(0, 7fr) minmax(0, 5fr); gap: 48px; align-items: center; padding: 40px 0 56px; } | |
| .px-new { display: inline-block; background: #FFD23F; font-family: "IBM Plex Mono", monospace; font-size: 12.5px; font-weight: 600; | |
| letter-spacing: .08em; text-transform: uppercase; padding: 6px 10px; } | |
| h1.px-h1 { font-size: clamp(44px, 6.2vw, 80px) !important; font-weight: 800 !important; line-height: .98 !important; | |
| letter-spacing: -.035em; margin: 22px 0 !important; } | |
| .px-h1 mark { background: linear-gradient(transparent 58%, #FFD23F 58% 92%, transparent 92%); color: inherit; padding: 0 .04em; } | |
| .px-lead { font-size: 20px; line-height: 1.5; color: #4B453E; margin: 0; max-width: 34em; } | |
| .px-cta { display: flex; gap: 12px; flex-wrap: wrap; margin-top: 30px; } | |
| .px-facts { display: grid; grid-template-columns: repeat(4, minmax(0, 1fr)); margin-top: 40px; border-top: 1.5px solid #161412; } | |
| .px-fact { padding: 14px 14px 0 0; display: flex; flex-direction: column; gap: 2px; } | |
| .px-fact b { font-size: 26px; font-weight: 800; letter-spacing: -.02em; } | |
| .px-fact span, .px-count, .px-msrc, .px-cap { font-family: "IBM Plex Mono", monospace; font-size: 12.5px; color: #6F675D; } | |
| .px-poster { background: #161412; aspect-ratio: 1 / 1; display: flex; flex-direction: column; justify-content: space-between; | |
| padding: 36px 36px 0; box-sizing: border-box; overflow: hidden; } | |
| .px-poster img { width: 100%; height: auto; display: block; } | |
| .px-poster .px-kicker { color: #C9BFAE !important; } | |
| .px-head { display: flex; align-items: flex-end; justify-content: space-between; gap: 20px; flex-wrap: wrap; margin: 72px 0 22px; } | |
| h2.px-h2 { font-size: clamp(32px, 4.4vw, 52px) !important; font-weight: 800 !important; line-height: 1 !important; | |
| letter-spacing: -.03em; margin: 10px 0 0 !important; } | |
| .px-work { gap: 0 !important; border: 1.5px solid #161412; background: #fff; } | |
| .px-panel { padding: 24px !important; gap: 16px !important; } | |
| .px-panel .form { background: transparent !important; border: 0 !important; gap: 16px !important; overflow: visible !important; } | |
| #px-result { flex-grow: 0 !important; } | |
| .px-panel .tabs { flex-grow: 1; } | |
| .px-panel + .px-panel { border-left: 1.5px solid #161412; } | |
| .counter { font-family: "IBM Plex Mono", monospace; font-size: 13px; color: #6F675D; text-align: right; margin-top: -6px; } | |
| .counter.over { color: #C02E15; font-weight: 600; } | |
| .px-beam .wrap:has(> label) { display: grid !important; grid-template-columns: repeat(3, minmax(0, 1fr)); gap: 0 !important; | |
| border: 1.5px solid #161412; } | |
| .px-beam label { justify-content: center; text-align: center; min-height: 48px; padding: 0 6px !important; font-weight: 600 !important; | |
| font-size: 15px !important; } | |
| .px-beam label + label { border-left: 1.5px solid #161412 !important; } | |
| .px-beam label:focus-within { outline: 2px solid #F2601C; outline-offset: -4px; } | |
| .px-beam label input { position: absolute; opacity: 0; width: 1px; height: 1px; } | |
| .px-btns { gap: 10px !important; } | |
| .px-btns button { min-height: 56px; font-size: 17px !important; } | |
| .px-row { display: flex; align-items: center; justify-content: space-between; gap: 12px; flex-wrap: wrap; min-height: 30px; } | |
| .stats { display: flex; flex-wrap: wrap; gap: 8px; } | |
| .chip { font-family: "IBM Plex Mono", monospace; font-size: 12.5px; font-weight: 600; padding: 5px 10px; background: #161412; color: #FBF6EC; } | |
| .chip.ok { background: #FFD23F; color: #161412; } | |
| .chip.n { background: #F3EBDB; color: #161412; } | |
| .diff { min-height: 240px; box-sizing: border-box; background: #FBF6EC; border: 1px solid #E2D8C6; padding: 16px 18px; | |
| font-size: 17px; line-height: 2.1; white-space: pre-wrap; } | |
| .diff.empty { color: #6F675D; line-height: 1.8; } | |
| .diff del { color: #C02E15; background: #FBE3DA; text-decoration: line-through; text-decoration-thickness: 2px; padding: 1px 5px; } | |
| .diff ins { color: #161412; background: #FFD23F; text-decoration: none; font-weight: 700; padding: 2px 6px; } | |
| #px-examples { margin-top: 40px; } | |
| #px-examples .gallery-item { border: 0 !important; border-bottom: 1px solid #E2D8C6 !important; background: transparent !important; | |
| text-align: left; min-height: 52px; } | |
| #px-examples .gallery-item:hover { color: #C02E15; } | |
| .px-model { box-shadow: 0 0 0 100vmax #F3EBDB; clip-path: inset(0 -100vmax); background: #F3EBDB; margin-top: 88px; padding: 72px 0 80px; } | |
| .px-sub { font-size: 19px; color: #4B453E; margin: 16px 0 0; max-width: 38em; } | |
| .px-metrics { display: grid; grid-template-columns: repeat(3, minmax(0, 1fr)); gap: 20px; margin-top: 40px; } | |
| .px-metric { background: #fff; border: 1.5px solid #161412; padding: 26px; display: flex; flex-direction: column; gap: 14px; } | |
| .px-big { font-size: 84px; font-weight: 800; line-height: .9; letter-spacing: -.04em; } | |
| .px-mlabel { font-size: 17px; font-weight: 700; margin: 0; } | |
| .px-msrc { margin: -8px 0 0; } | |
| .px-bars { display: flex; flex-direction: column; gap: 10px; margin-top: 6px; } | |
| .px-bar { display: grid; grid-template-columns: 58px minmax(0, 240px) auto; gap: 12px; align-items: center; | |
| font-family: "IBM Plex Mono", monospace; font-size: 12.5px; color: #4B453E; } | |
| .px-pbar { height: 14px; background: linear-gradient(#8F867A, #8F867A) 0 0 / calc(var(--n) * 12px) 100% no-repeat, #E6DCCB; PIXELS_BAR; } | |
| .px-pbar.s2 { background: linear-gradient(90deg, #FFD23F, #FF8A1F 50%, #D9381E) 0 0 / calc(var(--n) * 12px) 100% no-repeat, #E6DCCB; } | |
| .px-specs { display: flex; flex-direction: column; } | |
| .px-spec { display: flex; justify-content: space-between; align-items: baseline; gap: 12px; padding: 13px 0; border-bottom: 1px solid #E2D8C6; } | |
| .px-spec:last-child { border-bottom: 0; } | |
| .px-spec span { color: #4B453E; font-size: 15px; } | |
| .px-spec b { font-family: "IBM Plex Mono", monospace; font-weight: 600; font-size: 15px; } | |
| .px-cap { margin: 16px 0 0; } | |
| .px-section { padding-top: 72px; } | |
| .px-steps { display: grid; grid-template-columns: repeat(3, minmax(0, 1fr)); gap: 28px; margin-top: 30px; } | |
| .px-step { border-top: 1.5px solid #161412; padding-top: 18px; } | |
| .px-num { display: flex; gap: 2px; margin-bottom: 18px; } | |
| .px-num i { width: 14px; height: 14px; } | |
| .px-step h3 { font-size: 21px; font-weight: 800; letter-spacing: -.01em; margin: 0 0 8px; } | |
| .px-step p { margin: 0; color: #4B453E; font-size: 16px; } | |
| .px-note { margin-top: 56px; display: flex; gap: 12px; align-items: flex-start; background: #fff; border: 1.5px solid #161412; | |
| padding: 14px 18px; font-size: 15px; } | |
| .px-note svg { flex: none; margin-top: 2px; color: #F2601C; } | |
| .px-foot { box-shadow: 0 0 0 100vmax #161412; clip-path: inset(0 -100vmax); background: #161412; color: #FBF6EC; margin-top: 88px; } | |
| .px-foot-in { display: flex; justify-content: space-between; align-items: flex-end; gap: 28px; flex-wrap: wrap; padding: 48px 0 40px; } | |
| .px-foot img { width: min(100%, 360px); height: auto; display: block; } | |
| .px-foot nav { display: flex; gap: 24px; flex-wrap: wrap; font-size: 15px; } | |
| .px-foot a { color: #FBF6EC; } | |
| .px-foot a:hover { color: #FFB020; } | |
| @media (max-width: 860px) { | |
| .gradio-container { padding: 0 16px !important; } | |
| .px-hero, .px-metrics, .px-steps { grid-template-columns: minmax(0, 1fr); } | |
| .px-work { flex-direction: column; } | |
| .px-panel + .px-panel { border-left: 0; border-top: 1.5px solid #161412; } | |
| .px-facts { grid-template-columns: repeat(2, minmax(0, 1fr)); } | |
| .px-beam label { font-size: 13.5px !important; } | |
| .px-panel { padding: 16px !important; } | |
| .px-poster { padding: 24px 24px 0; } | |
| .px-stair.small { margin: 0 -24px; } | |
| .px-stair, .px-foot .px-stair { background-size: 60px 18px, 120px 18px, 180px 18px, 240px 18px, 300px 18px, 100% 18px; } | |
| } | |
| """.replace("STAIR_BIG", STAIR.format("200px", "400px", "600px", "800px", "1000px")) \ | |
| .replace("STAIR_SMALL", STAIR.format("80px", "160px", "240px", "320px", "100%")) \ | |
| .replace("STAIR_FOOT", STAIR.format("200px", "400px", "600px", "800px", "100%")) \ | |
| .replace("PIXELS_BIG", PIXELS.format(18, 20)).replace("PIXELS_BAR", PIXELS.format(10, 12)) | |
| HEAD = ('<link rel="preconnect" href="https://fonts.googleapis.com"><link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>' | |
| '<link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Archivo:wdth,wght@62..125,100..900' | |
| '&family=IBM+Plex+Mono:wght@400;500;600&display=swap">') | |
| INK, SAND, PAPER, LINE = "#161412", "#FBF6EC", "#FFFFFF", "#E2D8C6" | |
| THEME = gr.themes.Base( | |
| primary_hue=gr.themes.Color("#FFF8E0", "#FFEFB3", "#FFE27A", "#FFD23F", "#FFB020", "#FF8A1F", "#F2601C", "#D9381E", | |
| "#C02E15", "#8F2410", "#5E180B", name="glut"), | |
| neutral_hue=gr.themes.Color(SAND, "#F3EBDB", LINE, "#C9BFAE", "#8F867A", "#6F675D", "#4B453E", "#39342E", | |
| "#26221E", "#1C1916", INK, name="sand"), | |
| radius_size=gr.themes.sizes.radius_none, | |
| font=["Archivo", "ui-sans-serif", "system-ui", "sans-serif"], | |
| font_mono=["IBM Plex Mono", "ui-monospace", "monospace"], | |
| ).set( | |
| body_background_fill=SAND, body_text_color=INK, body_text_color_subdued="#4B453E", | |
| background_fill_primary=PAPER, background_fill_secondary=SAND, border_color_primary=LINE, border_color_accent=INK, | |
| block_background_fill="transparent", block_border_width="0px", block_shadow="none", block_padding="0px", | |
| block_title_text_color="#4B453E", block_title_background_fill="transparent", block_label_background_fill="transparent", | |
| block_title_padding="0px", block_label_padding="0px", block_label_text_color="#4B453E", block_info_text_color="#6F675D", | |
| input_background_fill=SAND, input_border_color=LINE, input_border_color_focus=INK, input_shadow="none", | |
| input_shadow_focus="none", input_placeholder_color="#6F675D", input_radius="0px", | |
| color_accent="#F2601C", color_accent_soft="#FFEFB3", link_text_color=INK, link_text_color_hover="#C02E15", | |
| checkbox_label_background_fill="transparent", checkbox_label_background_fill_hover="#F3EBDB", | |
| checkbox_label_background_fill_selected=INK, checkbox_label_text_color=INK, checkbox_label_text_color_selected=SAND, | |
| checkbox_label_border_color="transparent", checkbox_label_border_color_selected=INK, checkbox_label_shadow="none", | |
| button_border_width="1.5px", button_large_radius="0px", button_medium_radius="0px", button_small_radius="0px", | |
| button_primary_shadow="none", button_secondary_shadow="none", button_large_text_weight="700", button_medium_text_weight="700", | |
| button_primary_background_fill=INK, button_primary_background_fill_hover="#F2601C", button_primary_text_color=SAND, | |
| button_primary_text_color_hover=INK, button_primary_border_color=INK, button_primary_border_color_hover="#F2601C", | |
| button_secondary_background_fill="transparent", button_secondary_background_fill_hover=INK, | |
| button_secondary_text_color=INK, button_secondary_text_color_hover=SAND, button_secondary_border_color=INK, | |
| button_secondary_border_color_hover=INK, | |
| ) | |
| for k in [k for k in vars(THEME) if k.endswith("_dark")]: # ein Look, auch wenn der Browser Dunkelmodus will | |
| setattr(THEME, k, getattr(THEME, k[:-5], getattr(THEME, k))) | |
| EXAMPLES = [ | |
| "ich weis nicht das es so ist. Morgen gehen wir in die stadt , und kaufen ein.", | |
| "Die Studenten haben in der Biliothek einen intressanten Vortrag über umweltschutz gehört.", | |
| "Wen das Wetter schön ist gehen wir mit unseren freunden spaziren.", | |
| "Gestern habe ich meinen Freund getroffen der schon seid Jahren in Berlin Wohnt. " | |
| "Er hat mir erzählt das er bald ein neues Auto kaufen will, weil sein altes nicht mehr Zuverlässig ist.", | |
| ] | |
| with gr.Blocks(title="HORST – Deutsche Textkorrektur") as demo: | |
| gr.HTML(HERO, apply_default_css=False) | |
| with gr.Row(equal_height=True, elem_classes=["px-work"], elem_id="korrektur"): | |
| with gr.Column(scale=1, elem_classes=["px-panel"]): | |
| inp = gr.Textbox(lines=10, max_lines=20, label="Dein Text", | |
| placeholder=f"Füge hier einen deutschen Text ein, bis zu {MAX_WORDS} Wörter …") | |
| counter = gr.HTML(counter_html("")) | |
| beam = gr.Radio(choices=[("Schnell", 1), ("Ausgewogen", 3), ("Sehr genau", 5)], value=1, label="Genauigkeit", | |
| info="Höhere Genauigkeit braucht etwas mehr GPU-Zeit.", elem_classes=["px-beam"]) | |
| with gr.Row(elem_classes=["px-btns"]): | |
| clear_btn = gr.Button("Leeren", variant="secondary", scale=1) | |
| run_btn = gr.Button("Korrigieren →", variant="primary", scale=2) | |
| with gr.Column(scale=1, elem_classes=["px-panel"]): | |
| stats = gr.HTML(result_head(), apply_default_css=False, elem_id="px-result") | |
| with gr.Tabs(): | |
| with gr.Tab("Änderungen"): | |
| diff = gr.HTML(EMPTY_DIFF) | |
| with gr.Tab("Korrigierter Text"): | |
| out = gr.Textbox(lines=10, max_lines=20, label="Text zum Kopieren", interactive=False, buttons=["copy"]) | |
| gr.Examples(examples=[[e] for e in EXAMPLES], inputs=[inp], label="Beispiele zum Ausprobieren", elem_id="px-examples") | |
| gr.HTML(INFO, apply_default_css=False) | |
| inp.change(counter_html, inp, counter, show_progress="hidden", api_name=False) | |
| run_btn.click(correct, [inp, beam], out, api_name="correct").success(show_diff, [inp, out], [diff, stats], api_name=False) | |
| clear_btn.click(clear, None, [inp, out, diff, stats, counter], api_name=False) | |
| if __name__ == "__main__": | |
| demo.launch(theme=THEME, css=CSS, head=HEAD, favicon_path=str(HERE / "favicon.png")) | |