Per-prompt validation field

prompts.jsonl

@@ -1,15 +1,15 @@
-{"id": "tgb-t2i-03", "modality": "text-to-image", "style": "crayon, hand-drawn banner", "prompt": "Create a wide banner image (roughly 3:1) drawn in crayon on textured paper, like a bright, playful hand-drawn explainer. At the top left, a box labelled \"Boxty v1.0\" with a small subtitle \"tech vocabulary\", and next to it, in handwriting: \"WER 1.35%\". At the top right, a box labelled \"Jachnun v1.0\" with a small subtitle \"code-switching\", and next to it, in handwriting: \"WER 5.85%\". Between them, a third box containing a jumble of words: \"teudat zehut\", \"arnona\", \"makolet\", \"GitHub Gist\", \"Kubernetes\". From that middle box, one bold arrow points straight down to a larger box at the bottom labelled \"Sabich v2.2\", with handwritten next to it: \"Code-switching WER 4.91%\" and \"Tech WER 2.76%\". In the margin beside the arrow, a scribbled note with a little doodle arrow pointing at Sabich: \"accuracy preserved!\". No other text.", "input_image": null, "expected_text": ["Boxty v1.0", "tech vocabulary", "WER 1.35%", "Jachnun v1.0", "code-switching", "WER 5.85%", "teudat zehut", "arnona", "makolet", "GitHub Gist", "Kubernetes", "Sabich v2.2", "Code-switching WER 4.91%", "Tech WER 2.76%", "accuracy preserved!"], "structure_checks": ["wide banner aspect ratio", "Boxty top left, Jachnun top right, word box between them", "each WER sits next to the right model", "exactly one arrow, from the word box down to Sabich v2.2", "margin note beside the arrow, pointing at Sabich", "crayon texture throughout", "no extra text"], "notes": "WER figures: Boxty v1.0 tech WER from its model card (hf.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent, test revision f956dbb184c2); Jachnun v1.0 and Sabich v2.2 from the Sabich model card, English-Hebrew Code-Switching ASR Benchmark test set v2 (revision 48ec85a), checked 2026-10-28. Boxty's figure is on a different test set, so it is a label to reproduce, not a like-for-like comparison."}
-{"id": "tgb-t2i-04", "modality": "text-to-image", "style": "cover art, free style", "prompt": "Cover art for a blog post about fine-tuning a speech recognition model on one person's voice so it understands both tech jargon and Hebrew words dropped into English sentences. Two earlier specialist models (a tech-vocabulary model and a code-switching model) feed into one combined model that does both without getting worse at either. Show the idea visually: sound waves from a speaker turning into accurate text, two streams merging into one, a sense of things clicking into place. Warm, colourful and characterful, with a nod to Israeli street food (the models are named after foods: the combined one is Sabich, a stuffed pita). Leave clear space at the top for a title. Do not try to write a title or any sentences.", "input_image": null, "expected_text": [], "structure_checks": ["two streams visibly merge into one", "speech turning into text is legible as an idea", "clear space left at the top", "any text that appears is real, correctly spelled words; no pseudo-text or garbled glyphs"], "notes": "Descriptive prompt with no exact strings. text_exact passes only if the image contains no garbled or invented text."}
-{"id": "tgb-t2i-05", "modality": "text-to-image", "style": "photoreal animal banner", "prompt": "Wide 16:9 banner image for a GitHub repository of home-server backup scripts. Photorealistic, warm natural light, shallow depth of field. A three-toed sloth sits at a cluttered desk in a small server closet, one long claw resting on a keyboard, looking calmly pleased with itself. The monitor in front of it shows a progress bar with the text \"Backup 3 of 7 complete\" above it. Behind the sloth, a rack of servers with blinking green lights, and a handwritten paper sign taped to the rack reading \"DO NOT UNPLUG\". On the desk, a white mug printed with \"NAS\". No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["Backup 3 of 7 complete", "DO NOT UNPLUG", "NAS"], "structure_checks": ["photoreal sloth at a desk, claw on the keyboard", "progress bar on the monitor with the text above it", "paper sign taped to the server rack", "mug on the desk carries NAS", "no other text"], "notes": "Modelled on the author's sloth blog heroes (backupsloth, amazedslothgpts in Personal-Site-Blog)."}
-{"id": "tgb-t2i-06", "modality": "text-to-image", "style": "photoreal two-animal banner", "prompt": "Wide 16:9 banner image for a repository of command-line plugins for working with spreadsheets. Photorealistic, soft morning light from a kitchen window. A sloth and a capybara sit side by side at a wooden kitchen table, pair-programming on one open laptop. The sloth points at the screen with a long claw; the capybara looks sceptical. The laptop screen shows a dark terminal with a single line of large, clearly readable text: \"$ claude plugin install csv-tools\". Two coffee cups and a potted basil plant on the table. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["$ claude plugin install csv-tools"], "structure_checks": ["two photoreal animals: a sloth and a capybara", "sloth points at the screen, capybara looks sceptical", "terminal line readable on the laptop screen", "no other text"], "notes": "One exact string, but small and on a screen inside a photoreal scene: the hardest place for text in a banner."}
-{"id": "tgb-t2i-07", "modality": "text-to-image", "style": "mascot in a photoreal scene", "prompt": "Wide 16:9 banner image for a plugin that looks up bus arrival times. Photorealistic street photograph, late-afternoon sun, Jerusalem stone buildings. A small, cute, blocky terracotta-orange robot mascot (soft felt texture, square body, two dot eyes, stubby arms) stands at a bus stop, looking up at an electronic arrival board. The board is amber LED on black and shows exactly two rows, in this order: first row \"18  Central Station  4 min\", second row \"74  Malha Mall  11 min\". A city bus is pulling in behind the mascot. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["18", "Central Station", "4 min", "74", "Malha Mall", "11 min"], "structure_checks": ["mascot is a stylised character placed in a photoreal scene", "arrival board has exactly two rows", "row order: 18 first, 74 second, each with its own destination and time", "bus behind the mascot", "no other text (including on the bus)"], "notes": "Modelled on the author's Claude plugin banners (blocky mascot in a real Israeli street or airport). A real Jerusalem bus would carry Hebrew; the prompt forbids it so any lettering on the bus counts as invented text."}
-{"id": "tgb-t2i-08", "modality": "text-to-image", "style": "neon title banner", "prompt": "Wide 16:9 title banner for a GitHub README. A glowing neon sign mounted on a dark wall covered in tangled coloured cables and small circuit boards, retro arcade feel, with small pixel-art accents in the corners. The sign reads, on two lines, in a rounded bold typeface: first line \"Daniel's MCP\" glowing green-to-cyan, second line \"server collection\" glowing orange-to-pink. A small pixel-art sloth face sits centred above the text. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["Daniel's MCP", "server collection"], "structure_checks": ["two lines in the stated order", "line colours as asked (green-to-cyan, then orange-to-pink)", "pixel-art sloth face centred above", "no other text, including on the circuit boards"], "notes": "Modelled on the Claude-Code-Plugins README banner. The apostrophe counts toward text_exact."}
-{"id": "tgb-t2i-09", "modality": "text-to-image", "style": "photoreal animal hero, descriptive", "prompt": "Hero image for a blog post about running large language models on a GPU at home instead of paying for cloud APIs. Photorealistic and a little funny: a sloth in a cosy home office at night, lit by the glow of a large desktop PC with a big graphics card visible through its glass side panel, fans spinning. The sloth looks proud and slightly overheated. Through the window, a faint city skyline. Leave clear, uncluttered space on the left third for a title to be added later. Do not write any text, logos, labels or numbers anywhere in the image.", "input_image": null, "expected_text": [], "structure_checks": ["photoreal sloth with a home GPU rig", "left third left clear for a title", "any text that appears fails: no logos, brand names or pseudo-text on the PC or screen"], "notes": "Descriptive prompt with no exact strings. text_exact passes only if the image contains no text at all (PC hardware invites invented brand names)."}
-{"id": "tgb-i2i-01", "modality": "image-to-image", "style": "keep handwriting, add colour", "prompt": "Turn this whiteboard photo into a clean, colourful diagram. Remove the whiteboard frame, glare and background, and correct the perspective to a straight-on view on a white background. Keep my handwriting style but make it neater and easy to read. Keep every label, box and arrow exactly where it is and pointing the same way. Fix obvious misspellings of technical names (\"Proxmox\", \"Replicate\", \"Remote\") but do not add, remove or reword any label. Add colour: give each group of related boxes its own soft colour and make the arrows a contrasting colour.", "input_image": "inputs/whiteboard-mcp-01.jpg", "expected_text": ["WAN", "Telegram", "CF / Tailscale", "OpenClaw", "Docker", "VM", "META MCP", "META MCP 2", "Desktop Client", "No proxy possible", "LAN use", "LAN MCPs", "NAS", "HA", "SBCs", "Proxmox", "SSE Remote MCPs", "Replicate", "Google"], "structure_checks": ["Desktop Client -> META MCP 2", "Desktop Client -> VM", "WAN -> OpenClaw (labelled CF / Tailscale)", "META MCP -> LAN MCPs", "META MCP -> SSE Remote MCPs", "LAN MCPs -> NAS, HA, SBCs, Proxmox", "SSE Remote MCPs -> Replicate, Google", "upward arrow labelled LAN use", "handwriting character kept", "no labels added or dropped"]}
-{"id": "tgb-i2i-03", "modality": "image-to-image", "style": "chalkboard", "prompt": "Turn this whiteboard photo into a diagram drawn on a classic green chalkboard. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Use white chalk for the text and coloured chalk (yellow, pink, light blue) for the boxes and arrows, with subtle chalk dust and smudges. The lettering should be neat chalk handwriting. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-08.jpg", "expected_text": ["Few Shot Learning - 3", "Zoom Meeting Prep SOP", "Is my hair a mess?", "Y", "N", "COMB HAIR", "CHILL OUT!"], "structure_checks": ["question branches into Y on the left and N on the right", "Y -> COMB HAIR, N -> CHILL OUT!", "COMB HAIR and CHILL OUT! are boxed", "title underlined at top left", "green chalkboard look"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 08 (flowchart). expected_text from its human transcription; case-insensitive (the board mixes caps)."}
-{"id": "tgb-i2i-04", "modality": "image-to-image", "style": "sticky notes", "prompt": "Turn this whiteboard photo into a colourful sticky-note board. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Put it on a light cork-board background. Render each block of text and each box as its own sticky note (yellow, pink, blue, green) with a slight shadow and tilt, in a casual handwritten font, and draw the arrows as marker lines between the notes. Keep the little bot and smiley faces as small doodles. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-11.jpg", "expected_text": ["PERSONALISED AI", "BASIC IDEA", "Use-case:", "Human wants to create a task-specific AI agent", "& frontload context", "Tactic:", "Create AI agent to \"interview\" you", "transcribe", "(could be voice → voice!)", "extract content", "vector DB", "What's your fav food?", "PIZZA!", "CREATE MEMORY", "Daniel's fav food = PIZZA", "VECTOR DB"], "structure_checks": ["bot face -> speech bubble What's your fav food?", "PIZZA! reply next to the smiley face", "bot -> CREATE MEMORY box, PIZZA! -> memory box", "memory box -> VECTOR DB cylinder", "tactic line keeps its arrow chain in order", "bot and smiley faces kept"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 11 (diagram). 16 strings; case-insensitive. A dense board: tests whether splitting text across notes drops or rewords any of it."}
-{"id": "tgb-i2i-05", "modality": "image-to-image", "style": "comic book", "prompt": "Turn this whiteboard photo into a comic-book page. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Bold black ink outlines, halftone dot shading and a bright red, blue, yellow and green palette. Draw the smiley face as a cheerful human character and the box-headed face as a friendly robot, and put each line of dialogue in a speech bubble from the right character. Add action lines for energy, but no sound-effect words or any other extra text. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-13.jpg", "expected_text": ["PERSONALISED AI", "WORKFLOW 2", "Hey bot! Let's build out my food & drink context store!", "Sure! Let me see what gaps we have...", "READ", "vector DB", "Hmm. nothing about beer preferences!", "What beer do you prefer?", "Super bitter IPAs", "CREATE", "User's fav beer = bitter IPA"], "structure_checks": ["user lines come from the human, bot lines from the robot", "dialogue order unchanged top to bottom", "READ -> vector DB -> bot (Hmm...)", "CREATE arrow leads to the memory oval User's fav beer = bitter IPA", "no POW/ZAP or other invented words"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 13 (diagram). Case-insensitive. The comic style pulls towards sound-effect words, which count as invented text."}
-{"id": "tgb-i2i-06", "modality": "image-to-image", "style": "neon sign", "prompt": "Turn this whiteboard photo into a neon-sign installation. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Put it on a dark exposed-brick wall. Render every line, arrow and word as a glowing neon tube (pink, blue, white and yellow), with soft light bleeding onto the bricks; arrows are continuous bent tubes. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-21.jpg", "expected_text": ["Whiteboard", "Dev Spec", "Agent Dev", "Whiteboard(s)", "Gemini API", "Gemini", "Cleaned Tech Diagrams", "(image 2 image)", "Narrative of Agent Model", "(image 2 txt)", "Agent 2:", "Convert these things", "Plan def w/ human", "Generate tasks, define subagents,", "checkpoints etc"], "structure_checks": ["title chain Whiteboard -> Dev Spec -> Agent Dev", "Whiteboard(s) -> Gemini API", "Gemini forks to Cleaned Tech Diagrams and Narrative of Agent Model", "Convert these things -> Dev Spec", "two sub-bullets under Agent 2 in order"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 21 (diagram). Case-insensitive. Neon with this much text is the hardest style here for legibility."}
-{"id": "tgb-he-00", "modality": "text-to-image", "track": "hebrew", "style": "flat colour, one label", "prompt": "A single rounded box drawn in bright flat colours on a white background, like one element of a tech diagram. Inside the box, one Hebrew word in large clear type, written right to left exactly as given: \"פירגון\". Below the box, a smaller label in Hebrew: \"מילון מונחים\". No other text.", "input_image": null, "expected_text": ["פירגון", "מילון מונחים"], "structure_checks": ["one box", "Hebrew word inside the box, label below it", "letters in the right order (right to left), not mirrored", "final letters (ן, ם) in their final forms", "no extra text"], "notes": "Screening prompt for the Hebrew track: one generation per model. Only models that pass text_exact here go on to tgb-he-01 and tgb-he-02. פירגון is carried over from Hebrew-Image-Generation-Eval (2025-12), where only Gemini 3 Pro, Nano Banana Pro and Wan 2.5 rendered it."}
-{"id": "tgb-he-01", "modality": "text-to-image", "track": "hebrew", "style": "bold flat colour, Hebrew only", "prompt": "Create a bright, colourful flat-colour explainer diagram in Hebrew, laid out for a right-to-left reader. All text is Hebrew, written right to left, spelled exactly as given. Title at the top: \"איך עובד זיהוי דיבור\". Four boxes in a single row, reading from right to left, connected by arrows pointing left: \"הקלטה\" -> \"תמלול\" -> \"תיקון טקסט\" -> \"טקסט סופי\". So \"הקלטה\" is the rightmost box and \"טקסט סופי\" the leftmost. Below the row, a separate box labelled \"מילון מונחים\" with one arrow pointing up into \"תמלול\" only. Under \"טקסט סופי\", a small badge reading \"שיעור שגיאות 2.76%\". A small, friendly cartoon robot mascot on the right edge points at \"הקלטה\". At the bottom, a caption banner reading exactly: \"מודל קטן, דיוק גבוה\". No other text.", "input_image": null, "expected_text": ["איך עובד זיהוי דיבור", "הקלטה", "תמלול", "תיקון טקסט", "טקסט סופי", "מילון מונחים", "שיעור שגיאות 2.76%", "מודל קטן, דיוק גבוה"], "structure_checks": ["four boxes in one row in the stated right-to-left order", "three left-pointing arrows between them", "מילון מונחים sits below the row", "exactly one arrow from מילון מונחים, pointing up into תמלול", "badge under טקסט סופי", "character on the right edge pointing at הקלטה", "Hebrew letters not mirrored or reversed; final forms (ך, ם, ן) correct", "the percentage reads 2.76% (digits left to right, % in place)", "no extra text"], "notes": "Hebrew-only mirror of tgb-t2i-02. 2.76% is Sabich v2.2's tech WER (see tgb-t2i-03). Common failures to log in text_errors: reversed letter order, mirrored glyphs, non-final letter at word end, Latin or pseudo-Hebrew filler, digits reversed by the bidi algorithm (6.72%)."}
-{"id": "tgb-he-02", "modality": "text-to-image", "track": "hebrew", "style": "flat colour, mixed Hebrew and English", "prompt": "Create a tech diagram with clean bold lines and flat bright colours, mixing Hebrew and English text. Every string must be spelled exactly as given: Hebrew right to left, English left to right, and where they share a line each keeps its own direction. Title at the top: \"תמלול Hebrish\". Three boxes in a row, laid out right to left, with arrows pointing left: \"הקלטה\" (rightmost) -> \"Sabich v2.2\" -> \"Transcript\" (leftmost). Under \"Sabich v2.2\", a small subtitle: \"מודל מכוונן\". A small, friendly cartoon robot mascot on the right, next to \"הקלטה\", has a speech bubble saying exactly: \"יש לי meeting ב-Zoom מחר\". Under \"Transcript\", a sticky note reading: \"teudat zehut = תעודת זהות\". At the bottom, a caption banner: \"WER 4.91% ב-code-switching\". No other text.", "input_image": null, "expected_text": ["תמלול Hebrish", "הקלטה", "Sabich v2.2", "Transcript", "מודל מכוונן", "יש לי meeting ב-Zoom מחר", "teudat zehut = תעודת זהות", "WER 4.91% ב-code-switching"], "structure_checks": ["three boxes in one row in the stated right-to-left order", "two left-pointing arrows between them", "subtitle under Sabich v2.2", "speech bubble comes from the character beside הקלטה", "sticky note under Transcript", "in mixed strings each run keeps its own direction: English words not reversed, Hebrew words not reversed, word order in the sentence correct for a Hebrew reader", "Hebrew prefix ב- attached to Zoom and code-switching, not detached or moved", "no extra text"], "notes": "Code-switching subtask: Hebrew and English in the same string, as in Hebrish speech. 4.91% is Sabich v2.2's code-switching WER (see tgb-t2i-03). Grade word order inside mixed strings as a text error, not a structure error: 'מחר Zoom-ב meeting לי יש' is a fail."}
+{"id": "tgb-t2i-03", "modality": "text-to-image", "style": "crayon, hand-drawn banner", "prompt": "Create a wide banner image (roughly 3:1) drawn in crayon on textured paper, like a bright, playful hand-drawn explainer. At the top left, a box labelled \"Boxty v1.0\" with a small subtitle \"tech vocabulary\", and next to it, in handwriting: \"WER 1.35%\". At the top right, a box labelled \"Jachnun v1.0\" with a small subtitle \"code-switching\", and next to it, in handwriting: \"WER 5.85%\". Between them, a third box containing a jumble of words: \"teudat zehut\", \"arnona\", \"makolet\", \"GitHub Gist\", \"Kubernetes\". From that middle box, one bold arrow points straight down to a larger box at the bottom labelled \"Sabich v2.2\", with handwritten next to it: \"Code-switching WER 4.91%\" and \"Tech WER 2.76%\". In the margin beside the arrow, a scribbled note with a little doodle arrow pointing at Sabich: \"accuracy preserved!\". No other text.", "input_image": null, "expected_text": ["Boxty v1.0", "tech vocabulary", "WER 1.35%", "Jachnun v1.0", "code-switching", "WER 5.85%", "teudat zehut", "arnona", "makolet", "GitHub Gist", "Kubernetes", "Sabich v2.2", "Code-switching WER 4.91%", "Tech WER 2.76%", "accuracy preserved!"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["wide banner aspect ratio", "Boxty top left, Jachnun top right, word box between them", "each WER sits next to the right model", "exactly one arrow, from the word box down to Sabich v2.2", "margin note beside the arrow, pointing at Sabich", "crayon texture throughout", "no extra text"], "notes": "WER figures: Boxty v1.0 tech WER from its model card (hf.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent, test revision f956dbb184c2); Jachnun v1.0 and Sabich v2.2 from the Sabich model card, English-Hebrew Code-Switching ASR Benchmark test set v2 (revision 48ec85a), checked 2026-10-28. Boxty's figure is on a different test set, so it is a label to reproduce, not a like-for-like comparison."}
+{"id": "tgb-t2i-04", "modality": "text-to-image", "style": "cover art, free style", "prompt": "Cover art for a blog post about fine-tuning a speech recognition model on one person's voice so it understands both tech jargon and Hebrew words dropped into English sentences. Two earlier specialist models (a tech-vocabulary model and a code-switching model) feed into one combined model that does both without getting worse at either. Show the idea visually: sound waves from a speaker turning into accurate text, two streams merging into one, a sense of things clicking into place. Warm, colourful and characterful, with a nod to Israeli street food (the models are named after foods: the combined one is Sabich, a stuffed pita). Leave clear space at the top for a title. Do not try to write a title or any sentences.", "input_image": null, "expected_text": [], "validation": [], "structure_checks": ["two streams visibly merge into one", "speech turning into text is legible as an idea", "clear space left at the top", "any text that appears is real, correctly spelled words; no pseudo-text or garbled glyphs"], "notes": "Descriptive prompt with no exact strings, no title and no diagram elements: ratings only, no validation ticks. structure_checks are a checklist for the ratings."}
+{"id": "tgb-t2i-05", "modality": "text-to-image", "style": "photoreal animal banner", "prompt": "Wide 16:9 banner image for a GitHub repository of home-server backup scripts. Photorealistic, warm natural light, shallow depth of field. A three-toed sloth sits at a cluttered desk in a small server closet, one long claw resting on a keyboard, looking calmly pleased with itself. The monitor in front of it shows a progress bar with the text \"Backup 3 of 7 complete\" above it. Behind the sloth, a rack of servers with blinking green lights, and a handwritten paper sign taped to the rack reading \"DO NOT UNPLUG\". On the desk, a white mug printed with \"NAS\". No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["Backup 3 of 7 complete", "DO NOT UNPLUG", "NAS"], "validation": ["text_exact", "text_legible"], "structure_checks": ["photoreal sloth at a desk, claw on the keyboard", "progress bar on the monitor with the text above it", "paper sign taped to the server rack", "mug on the desk carries NAS", "no other text"], "notes": "Modelled on the author's sloth blog heroes (backupsloth, amazedslothgpts in Personal-Site-Blog)."}
+{"id": "tgb-t2i-06", "modality": "text-to-image", "style": "photoreal two-animal banner", "prompt": "Wide 16:9 banner image for a repository of command-line plugins for working with spreadsheets. Photorealistic, soft morning light from a kitchen window. A sloth and a capybara sit side by side at a wooden kitchen table, pair-programming on one open laptop. The sloth points at the screen with a long claw; the capybara looks sceptical. The laptop screen shows a dark terminal with a single line of large, clearly readable text: \"$ claude plugin install csv-tools\". Two coffee cups and a potted basil plant on the table. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["$ claude plugin install csv-tools"], "validation": ["text_exact", "text_legible"], "structure_checks": ["two photoreal animals: a sloth and a capybara", "sloth points at the screen, capybara looks sceptical", "terminal line readable on the laptop screen", "no other text"], "notes": "One exact string, but small and on a screen inside a photoreal scene: the hardest place for text in a banner."}
+{"id": "tgb-t2i-07", "modality": "text-to-image", "style": "mascot in a photoreal scene", "prompt": "Wide 16:9 banner image for a plugin that looks up bus arrival times. Photorealistic street photograph, late-afternoon sun, Jerusalem stone buildings. A small, cute, blocky terracotta-orange robot mascot (soft felt texture, square body, two dot eyes, stubby arms) stands at a bus stop, looking up at an electronic arrival board. The board is amber LED on black and shows exactly two rows, in this order: first row \"18  Central Station  4 min\", second row \"74  Malha Mall  11 min\". A city bus is pulling in behind the mascot. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["18", "Central Station", "4 min", "74", "Malha Mall", "11 min"], "validation": ["text_exact", "text_legible"], "structure_checks": ["mascot is a stylised character placed in a photoreal scene", "arrival board has exactly two rows", "row order: 18 first, 74 second, each with its own destination and time", "bus behind the mascot", "no other text (including on the bus)"], "notes": "Modelled on the author's Claude plugin banners (blocky mascot in a real Israeli street or airport). A real Jerusalem bus would carry Hebrew; the prompt forbids it so any lettering on the bus counts as invented text."}
+{"id": "tgb-t2i-08", "modality": "text-to-image", "style": "neon title banner", "prompt": "Wide 16:9 title banner for a GitHub README. A glowing neon sign mounted on a dark wall covered in tangled coloured cables and small circuit boards, retro arcade feel, with small pixel-art accents in the corners. The sign reads, on two lines, in a rounded bold typeface: first line \"Daniel's MCP\" glowing green-to-cyan, second line \"server collection\" glowing orange-to-pink. A small pixel-art sloth face sits centred above the text. No other text, logos or lettering anywhere in the image.", "input_image": null, "expected_text": ["Daniel's MCP", "server collection"], "validation": ["text_exact", "text_legible"], "structure_checks": ["two lines in the stated order", "line colours as asked (green-to-cyan, then orange-to-pink)", "pixel-art sloth face centred above", "no other text, including on the circuit boards"], "notes": "Modelled on the Claude-Code-Plugins README banner. The apostrophe counts toward text_exact."}
+{"id": "tgb-t2i-09", "modality": "text-to-image", "style": "photoreal animal hero, descriptive", "prompt": "Hero image for a blog post about running large language models on a GPU at home instead of paying for cloud APIs. Photorealistic and a little funny: a sloth in a cosy home office at night, lit by the glow of a large desktop PC with a big graphics card visible through its glass side panel, fans spinning. The sloth looks proud and slightly overheated. Through the window, a faint city skyline. Leave clear, uncluttered space on the left third for a title to be added later. Do not write any text, logos, labels or numbers anywhere in the image.", "input_image": null, "expected_text": [], "validation": [], "structure_checks": ["photoreal sloth with a home GPU rig", "left third left clear for a title", "any text that appears fails: no logos, brand names or pseudo-text on the PC or screen"], "notes": "Descriptive prompt with no exact strings, no title and no diagram elements: ratings only, no validation ticks. Invented text or brand names on the PC count against the ratings."}
+{"id": "tgb-i2i-01", "modality": "image-to-image", "style": "keep handwriting, add colour", "prompt": "Turn this whiteboard photo into a clean, colourful diagram. Remove the whiteboard frame, glare and background, and correct the perspective to a straight-on view on a white background. Keep my handwriting style but make it neater and easy to read. Keep every label, box and arrow exactly where it is and pointing the same way. Fix obvious misspellings of technical names (\"Proxmox\", \"Replicate\", \"Remote\") but do not add, remove or reword any label. Add colour: give each group of related boxes its own soft colour and make the arrows a contrasting colour.", "input_image": "inputs/whiteboard-mcp-01.jpg", "expected_text": ["WAN", "Telegram", "CF / Tailscale", "OpenClaw", "Docker", "VM", "META MCP", "META MCP 2", "Desktop Client", "No proxy possible", "LAN use", "LAN MCPs", "NAS", "HA", "SBCs", "Proxmox", "SSE Remote MCPs", "Replicate", "Google"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["Desktop Client -> META MCP 2", "Desktop Client -> VM", "WAN -> OpenClaw (labelled CF / Tailscale)", "META MCP -> LAN MCPs", "META MCP -> SSE Remote MCPs", "LAN MCPs -> NAS, HA, SBCs, Proxmox", "SSE Remote MCPs -> Replicate, Google", "upward arrow labelled LAN use", "handwriting character kept", "no labels added or dropped"]}
+{"id": "tgb-i2i-03", "modality": "image-to-image", "style": "chalkboard", "prompt": "Turn this whiteboard photo into a diagram drawn on a classic green chalkboard. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Use white chalk for the text and coloured chalk (yellow, pink, light blue) for the boxes and arrows, with subtle chalk dust and smudges. The lettering should be neat chalk handwriting. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-08.jpg", "expected_text": ["Few Shot Learning - 3", "Zoom Meeting Prep SOP", "Is my hair a mess?", "Y", "N", "COMB HAIR", "CHILL OUT!"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["question branches into Y on the left and N on the right", "Y -> COMB HAIR, N -> CHILL OUT!", "COMB HAIR and CHILL OUT! are boxed", "title underlined at top left", "green chalkboard look"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 08 (flowchart). expected_text from its human transcription; case-insensitive (the board mixes caps)."}
+{"id": "tgb-i2i-04", "modality": "image-to-image", "style": "sticky notes", "prompt": "Turn this whiteboard photo into a colourful sticky-note board. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Put it on a light cork-board background. Render each block of text and each box as its own sticky note (yellow, pink, blue, green) with a slight shadow and tilt, in a casual handwritten font, and draw the arrows as marker lines between the notes. Keep the little bot and smiley faces as small doodles. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-11.jpg", "expected_text": ["PERSONALISED AI", "BASIC IDEA", "Use-case:", "Human wants to create a task-specific AI agent", "& frontload context", "Tactic:", "Create AI agent to \"interview\" you", "transcribe", "(could be voice → voice!)", "extract content", "vector DB", "What's your fav food?", "PIZZA!", "CREATE MEMORY", "Daniel's fav food = PIZZA", "VECTOR DB"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["bot face -> speech bubble What's your fav food?", "PIZZA! reply next to the smiley face", "bot -> CREATE MEMORY box, PIZZA! -> memory box", "memory box -> VECTOR DB cylinder", "tactic line keeps its arrow chain in order", "bot and smiley faces kept"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 11 (diagram). 16 strings; case-insensitive. A dense board: tests whether splitting text across notes drops or rewords any of it."}
+{"id": "tgb-i2i-05", "modality": "image-to-image", "style": "comic book", "prompt": "Turn this whiteboard photo into a comic-book page. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Bold black ink outlines, halftone dot shading and a bright red, blue, yellow and green palette. Draw the smiley face as a cheerful human character and the box-headed face as a friendly robot, and put each line of dialogue in a speech bubble from the right character. Add action lines for energy, but no sound-effect words or any other extra text. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-13.jpg", "expected_text": ["PERSONALISED AI", "WORKFLOW 2", "Hey bot! Let's build out my food & drink context store!", "Sure! Let me see what gaps we have...", "READ", "vector DB", "Hmm. nothing about beer preferences!", "What beer do you prefer?", "Super bitter IPAs", "CREATE", "User's fav beer = bitter IPA"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["user lines come from the human, bot lines from the robot", "dialogue order unchanged top to bottom", "READ -> vector DB -> bot (Hmm...)", "CREATE arrow leads to the memory oval User's fav beer = bitter IPA", "no POW/ZAP or other invented words"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 13 (diagram). Case-insensitive. The comic style pulls towards sound-effect words, which count as invented text."}
+{"id": "tgb-i2i-06", "modality": "image-to-image", "style": "neon sign", "prompt": "Turn this whiteboard photo into a neon-sign installation. Remove the physical whiteboard, its frame, the pen tray, glare and the wall. Correct the perspective so the result is a straight-on view. Put it on a dark exposed-brick wall. Render every line, arrow and word as a glowing neon tube (pink, blue, white and yellow), with soft light bleeding onto the bricks; arrows are continuous bent tubes. Preserve every word, box and arrow, and do not add, remove or reword any text: copy each word as written, correcting only what the handwriting makes ambiguous. Keep every arrow pointing the same way between the same items.", "input_image": "inputs/whiteboards-21.jpg", "expected_text": ["Whiteboard", "Dev Spec", "Agent Dev", "Whiteboard(s)", "Gemini API", "Gemini", "Cleaned Tech Diagrams", "(image 2 image)", "Narrative of Agent Model", "(image 2 txt)", "Agent 2:", "Convert these things", "Plan def w/ human", "Generate tasks, define subagents,", "checkpoints etc"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["title chain Whiteboard -> Dev Spec -> Agent Dev", "Whiteboard(s) -> Gemini API", "Gemini forks to Cleaned Tech Diagrams and Narrative of Agent Model", "Convert these things -> Dev Spec", "two sub-bullets under Agent 2 in order"], "notes": "Input: https://huggingface.co/datasets/danielrosehill/Whiteboards sample 21 (diagram). Case-insensitive. Neon with this much text is the hardest style here for legibility."}
+{"id": "tgb-he-00", "modality": "text-to-image", "track": "hebrew", "style": "flat colour, one label", "prompt": "A single rounded box drawn in bright flat colours on a white background, like one element of a tech diagram. Inside the box, one Hebrew word in large clear type, written right to left exactly as given: \"פירגון\". Below the box, a smaller label in Hebrew: \"מילון מונחים\". No other text.", "input_image": null, "expected_text": ["פירגון", "מילון מונחים"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["one box", "Hebrew word inside the box, label below it", "letters in the right order (right to left), not mirrored", "final letters (ן, ם) in their final forms", "no extra text"], "notes": "Screening prompt for the Hebrew track: one generation per model. Only models that pass text_exact here go on to tgb-he-01 and tgb-he-02. פירגון is carried over from Hebrew-Image-Generation-Eval (2025-12), where only Gemini 3 Pro, Nano Banana Pro and Wan 2.5 rendered it."}
+{"id": "tgb-he-01", "modality": "text-to-image", "track": "hebrew", "style": "bold flat colour, Hebrew only", "prompt": "Create a bright, colourful flat-colour explainer diagram in Hebrew, laid out for a right-to-left reader. All text is Hebrew, written right to left, spelled exactly as given. Title at the top: \"איך עובד זיהוי דיבור\". Four boxes in a single row, reading from right to left, connected by arrows pointing left: \"הקלטה\" -> \"תמלול\" -> \"תיקון טקסט\" -> \"טקסט סופי\". So \"הקלטה\" is the rightmost box and \"טקסט סופי\" the leftmost. Below the row, a separate box labelled \"מילון מונחים\" with one arrow pointing up into \"תמלול\" only. Under \"טקסט סופי\", a small badge reading \"שיעור שגיאות 2.76%\". A small, friendly cartoon robot mascot on the right edge points at \"הקלטה\". At the bottom, a caption banner reading exactly: \"מודל קטן, דיוק גבוה\". No other text.", "input_image": null, "expected_text": ["איך עובד זיהוי דיבור", "הקלטה", "תמלול", "תיקון טקסט", "טקסט סופי", "מילון מונחים", "שיעור שגיאות 2.76%", "מודל קטן, דיוק גבוה"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["four boxes in one row in the stated right-to-left order", "three left-pointing arrows between them", "מילון מונחים sits below the row", "exactly one arrow from מילון מונחים, pointing up into תמלול", "badge under טקסט סופי", "character on the right edge pointing at הקלטה", "Hebrew letters not mirrored or reversed; final forms (ך, ם, ן) correct", "the percentage reads 2.76% (digits left to right, % in place)", "no extra text"], "notes": "Hebrew-only mirror of tgb-t2i-02. 2.76% is Sabich v2.2's tech WER (see tgb-t2i-03). Common failures to log in text_errors: reversed letter order, mirrored glyphs, non-final letter at word end, Latin or pseudo-Hebrew filler, digits reversed by the bidi algorithm (6.72%)."}
+{"id": "tgb-he-02", "modality": "text-to-image", "track": "hebrew", "style": "flat colour, mixed Hebrew and English", "prompt": "Create a tech diagram with clean bold lines and flat bright colours, mixing Hebrew and English text. Every string must be spelled exactly as given: Hebrew right to left, English left to right, and where they share a line each keeps its own direction. Title at the top: \"תמלול Hebrish\". Three boxes in a row, laid out right to left, with arrows pointing left: \"הקלטה\" (rightmost) -> \"Sabich v2.2\" -> \"Transcript\" (leftmost). Under \"Sabich v2.2\", a small subtitle: \"מודל מכוונן\". A small, friendly cartoon robot mascot on the right, next to \"הקלטה\", has a speech bubble saying exactly: \"יש לי meeting ב-Zoom מחר\". Under \"Transcript\", a sticky note reading: \"teudat zehut = תעודת זהות\". At the bottom, a caption banner: \"WER 4.91% ב-code-switching\". No other text.", "input_image": null, "expected_text": ["תמלול Hebrish", "הקלטה", "Sabich v2.2", "Transcript", "מודל מכוונן", "יש לי meeting ב-Zoom מחר", "teudat zehut = תעודת זהות", "WER 4.91% ב-code-switching"], "validation": ["text_exact", "text_legible", "structure_ok", "meaning_preserved"], "structure_checks": ["three boxes in one row in the stated right-to-left order", "two left-pointing arrows between them", "subtitle under Sabich v2.2", "speech bubble comes from the character beside הקלטה", "sticky note under Transcript", "in mixed strings each run keeps its own direction: English words not reversed, Hebrew words not reversed, word order in the sentence correct for a Hebrew reader", "Hebrew prefix ב- attached to Zoom and code-switching, not detached or moved", "no extra text"], "notes": "Code-switching subtask: Hebrew and English in the same string, as in Hebrish speech. 4.91% is Sabich v2.2's code-switching WER (see tgb-t2i-03). Grade word order inside mixed strings as a text error, not a structure error: 'מחר Zoom-ב meeting לי יש' is a fail."}