{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"学习路径： 图像怎样约束句子与回答","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html","markdown":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.md","pages":[{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","unit":"学习路径： 图像怎样约束句子与回答","unitId":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","chapter":"course","number":1}],"aliases":[{"oldId":"b-7ed12842c8bb768f852ae6a1","targetId":"b-7ed12842c8bb768f852ae6a1","page":"reading-opening","operation":"move","beforeSHA256":"926fc2429c1a5c837765e5152b3b61fda31c10537785a439d49dec46e82a8536","afterSHA256":"926fc2429c1a5c837765e5152b3b61fda31c10537785a439d49dec46e82a8536","reason":"Same source body in authored semantic page"},{"oldId":"b-f0ffd1f863e8487a62a9ba6f","targetId":"b-f0ffd1f863e8487a62a9ba6f","page":"reading-opening","operation":"move","beforeSHA256":"e5df180ed030b50c84a576c8911ba2a69c96ab192fcfac5e7686b63ab5af678a","afterSHA256":"e5df180ed030b50c84a576c8911ba2a69c96ab192fcfac5e7686b63ab5af678a","reason":"Same source body in authored semantic page"},{"oldId":"b-bbd330f22eb564c2f86a7cae","targetId":"b-bbd330f22eb564c2f86a7cae","page":"reading-opening","operation":"move","beforeSHA256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","afterSHA256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","reason":"Same source body in authored semantic page"},{"oldId":"b-d18bdbde03356fe93787e87b","targetId":"b-d18bdbde03356fe93787e87b","page":"reading-opening","operation":"move","beforeSHA256":"5790b0afb52508f0174b1bc581f7847035a51c5c2a46e5c821bcd9d5d61bcbb0","afterSHA256":"5790b0afb52508f0174b1bc581f7847035a51c5c2a46e5c821bcd9d5d61bcbb0","reason":"Same source body in authored semantic page"},{"oldId":"b-25dcd81eb023db10eb3a2c88","targetId":"b-25dcd81eb023db10eb3a2c88","page":"reading-opening","operation":"move","beforeSHA256":"a57513e59eb07960710e5ece7951cc57af4a94c446630eb5fc6d70b460982042","afterSHA256":"a57513e59eb07960710e5ece7951cc57af4a94c446630eb5fc6d70b460982042","reason":"Same source body in authored semantic page"},{"oldId":"b-9afc0453e7e98eee80706045","targetId":"b-9afc0453e7e98eee80706045","page":"reading-opening","operation":"move","beforeSHA256":"54bacb112e17b840a5a310ce968bdae21f53376a4ba54a33e651576ea0a41a11","afterSHA256":"54bacb112e17b840a5a310ce968bdae21f53376a4ba54a33e651576ea0a41a11","reason":"Same source body in authored semantic page"},{"oldId":"b-d3f1be25db81da7f2757dd5e","targetId":"b-d3f1be25db81da7f2757dd5e","page":"reading-opening","operation":"move","beforeSHA256":"a39bb3bfff571024e21cbf171c189671131c82900779f89b74b1609b897f31b1","afterSHA256":"a39bb3bfff571024e21cbf171c189671131c82900779f89b74b1609b897f31b1","reason":"Same source body in authored semantic page"},{"oldId":"b-ee9bf710749bbfe0b28aca01","targetId":"b-ee9bf710749bbfe0b28aca01","page":"reading-opening","operation":"move","beforeSHA256":"4790c4b243b6afc33a3b0a94b904dcef5a39431db1ef3a267d3dec7d13a4df01","afterSHA256":"4790c4b243b6afc33a3b0a94b904dcef5a39431db1ef3a267d3dec7d13a4df01","reason":"Same source body in authored semantic page"},{"oldId":"b-062a2db341a1bfd00a3d5880","targetId":"b-062a2db341a1bfd00a3d5880","page":"reading-opening","operation":"move","beforeSHA256":"a2c341314db4e1d785717eeb1afabfbccce760d49244c5ce15bae05cabd5117a","afterSHA256":"9a2655d1601e5f06db61c1f983135287b33ee7b26e382647125686614f100163","reason":"Same source body in authored semantic page"},{"oldId":"b-941ff9fcdd15ded41ce68a78","targetId":"b-941ff9fcdd15ded41ce68a78","page":"reading-opening","operation":"move","beforeSHA256":"36843086e2b941e4d4724082772339fb7dc1f03fc1e5596c7f355166ed31e988","afterSHA256":"36843086e2b941e4d4724082772339fb7dc1f03fc1e5596c7f355166ed31e988","reason":"Same source body in authored semantic page"}],"blocks":[{"id":"b-7ed12842c8bb768f852ae6a1","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-7ed12842c8bb768f852ae6a1","content":"[LAB 14 · 图像问答：视觉条件与问题先验](/course/notebooks/11d9f74871544a77b42024a9fe06b81a.html)","sha256":"926fc2429c1a5c837765e5152b3b61fda31c10537785a439d49dec46e82a8536","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-f0ffd1f863e8487a62a9ba6f","prerequisites":[],"relatedBlocks":[]},{"id":"b-f0ffd1f863e8487a62a9ba6f","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-f0ffd1f863e8487a62a9ba6f","content":"# 学习路径： 图像怎样约束句子与回答","sha256":"e5df180ed030b50c84a576c8911ba2a69c96ab192fcfac5e7686b63ab5af678a","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-7ed12842c8bb768f852ae6a1","nextBlock":"b-bbd330f22eb564c2f86a7cae","prerequisites":[{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-bbd330f22eb564c2f86a7cae","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-bbd330f22eb564c2f86a7cae","content":"按问题继续","sha256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f0ffd1f863e8487a62a9ba6f","nextBlock":"b-d18bdbde03356fe93787e87b","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d18bdbde03356fe93787e87b","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-d18bdbde03356fe93787e87b","content":"RGB+问题输入、question-only对照、同图换问与同问换图","sha256":"5790b0afb52508f0174b1bc581f7847035a51c5c2a46e5c821bcd9d5d61bcbb0","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-bbd330f22eb564c2f86a7cae","nextBlock":"b-25dcd81eb023db10eb3a2c88","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-25dcd81eb023db10eb3a2c88","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-25dcd81eb023db10eb3a2c88","content":"**学习路径：** [图像怎样约束句子与回答](/course/study/caption-vqa.html)","sha256":"a57513e59eb07960710e5ece7951cc57af4a94c446630eb5fc6d70b460982042","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-d18bdbde03356fe93787e87b","nextBlock":"b-9afc0453e7e98eee80706045","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-9afc0453e7e98eee80706045","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-9afc0453e7e98eee80706045","content":"同图换问题，模型读什么会变；同问换图片，答案也应跟着变。先读[P54](/course/pages/l3-r2-p54.html)，再看第2节按图切分和第4节逐问题结果。","sha256":"54bacb112e17b840a5a310ce968bdae21f53376a4ba54a33e651576ea0a41a11","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-25dcd81eb023db10eb3a2c88","nextBlock":"b-d3f1be25db81da7f2757dd5e","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d3f1be25db81da7f2757dd5e","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-d3f1be25db81da7f2757dd5e","content":"开头字典例是已知属性oracle，主体才是读取RGB像素与问题ID的CNN，并单独训练question-only。三类问题为数量、左色、右形，答案为有限词表；这不等于开放语言问答。[完整控制和证据](/course/blocks/vqa-pixel-evidence.html)。","sha256":"a39bb3bfff571024e21cbf171c189671131c82900779f89b74b1609b897f31b1","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-9afc0453e7e98eee80706045","nextBlock":"b-ee9bf710749bbfe0b28aca01","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-ee9bf710749bbfe0b28aca01","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-ee9bf710749bbfe0b28aca01","content":"读完应能解释question-only为何约44.44%，并从同图换问/同问换图各拿一个例子支持“模型使用了条件”。只读可先看2026-09-22保存结果，运行则按全部代码顺序；本次整理不改历史成绩。若答案需要一个区域而非类别，继续[LAB 15](https://deepnote.com/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/7dc83786d067465aa76264475db2c62d)。","sha256":"4790c4b243b6afc33a3b0a94b904dcef5a39431db1ef3a267d3dec7d13a4df01","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-d3f1be25db81da7f2757dd5e","nextBlock":"b-062a2db341a1bfd00a3d5880","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-062a2db341a1bfd00a3d5880","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-062a2db341a1bfd00a3d5880","content":"![同图换问与同问换图](/assets/display/2f7b2168f5b21ee2095709fa.webp)","sha256":"9a2655d1601e5f06db61c1f983135287b33ee7b26e382647125686614f100163","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ee9bf710749bbfe0b28aca01","nextBlock":"b-941ff9fcdd15ded41ce68a78","prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-941ff9fcdd15ded41ce68a78","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/9efe941652bc4525b09d4348dcef8323.html#b-941ff9fcdd15ded41ce68a78","content":"前两幅共享形状、位置和随机噪声，只改左侧颜色；第三幅只用于展示另一个问题条件，不是与前两幅只改一项。想让答案直接成为一个区域，继续 [LAB 15 · 语言指代小练习：目标、参照与关系](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/7dc83786d067465aa76264475db2c62d?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=7dc83786d067465aa76264475db2c62d&utm_term=get_notebook)。","sha256":"36843086e2b941e4d4724082772339fb7dc1f03fc1e5596c7f355166ed31e988","chapter":{"id":"course","title":"学习路径： 图像怎样约束句子与回答"},"unit":{"id":"notebook-blocks/9efe941652bc4525b09d4348dcef8323","title":"学习路径： 图像怎样约束句子与回答"},"page":{"id":"reading-opening","title":"学习路径： 图像怎样约束句子与回答","number":1},"sectionTitle":"学习路径： 图像怎样约束句子与回答","currentQuestion":"学习路径： 图像怎样约束句子与回答","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-062a2db341a1bfd00a3d5880","nextBlock":null,"prerequisites":[{"id":"b-f0ffd1f863e8487a62a9ba6f","url":"#b-f0ffd1f863e8487a62a9ba6f","role":"section-title"},{"id":"b-7ed12842c8bb768f852ae6a1","url":"#b-7ed12842c8bb768f852ae6a1","role":"problem-context"}],"relatedBlocks":[]}]}