{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"学习路径： 文字候选与图文检索","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html","markdown":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.md","pages":[{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","unit":"学习路径： 文字候选与图文检索","unitId":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","chapter":"course","number":1}],"aliases":[{"oldId":"b-8e7e4a754c96272f85ed9ab1","targetId":"b-8e7e4a754c96272f85ed9ab1","page":"reading-opening","operation":"move","beforeSHA256":"343fb71807f077069b41c6495a8824ea4d353c740175b5775ccc5145b3bf19ca","afterSHA256":"343fb71807f077069b41c6495a8824ea4d353c740175b5775ccc5145b3bf19ca","reason":"Same source body in authored semantic page"},{"oldId":"b-4cafcedac2a9055b7b5a9009","targetId":"b-4cafcedac2a9055b7b5a9009","page":"reading-opening","operation":"move","beforeSHA256":"cf5bb768e672561b711388225dafb192390defe7ee096fd0c43eb07b12326246","afterSHA256":"cf5bb768e672561b711388225dafb192390defe7ee096fd0c43eb07b12326246","reason":"Same source body in authored semantic page"},{"oldId":"b-663d90a41011b1e6eace0423","targetId":"b-663d90a41011b1e6eace0423","page":"reading-opening","operation":"move","beforeSHA256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","afterSHA256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","reason":"Same source body in authored semantic page"},{"oldId":"b-bf239ab282cf8ee448fe420b","targetId":"b-bf239ab282cf8ee448fe420b","page":"reading-opening","operation":"move","beforeSHA256":"41c70c2c07b82678ff0a60f2d009b6afa750c5c691d109e754a69abf23868b50","afterSHA256":"41c70c2c07b82678ff0a60f2d009b6afa750c5c691d109e754a69abf23868b50","reason":"Same source body in authored semantic page"},{"oldId":"b-3db83c6216a1320914e49251","targetId":"b-3db83c6216a1320914e49251","page":"reading-opening","operation":"move","beforeSHA256":"bab4bd804571aa5afbaf197e9c1539510fc766578c9825aa46cadbb18444cc32","afterSHA256":"bab4bd804571aa5afbaf197e9c1539510fc766578c9825aa46cadbb18444cc32","reason":"Same source body in authored semantic page"},{"oldId":"b-7ad29d8c288e51bfb53bf83f","targetId":"b-7ad29d8c288e51bfb53bf83f","page":"reading-opening","operation":"move","beforeSHA256":"f727b99f9d6274f0111bb10a0383f86a10271b27674851de4ca88872fc35141a","afterSHA256":"f727b99f9d6274f0111bb10a0383f86a10271b27674851de4ca88872fc35141a","reason":"Same source body in authored semantic page"},{"oldId":"b-edac41e4bc5c95ad18a03ea9","targetId":"b-edac41e4bc5c95ad18a03ea9","page":"reading-opening","operation":"rewrite","beforeSHA256":"68534496237ab167741e9fd0e94e2d15f431f288da2adff9cc1c9716ff8ef303","afterSHA256":"1d235867f8149bef480564be89cbe45135e146a1a5ca2cd3ac79ed658b91718b","reason":"Same source identity with figure semantics or reading context enriched"},{"oldId":"b-f15d6e975fbb4775abb90120","targetId":"b-f15d6e975fbb4775abb90120","page":"reading-opening","operation":"move","beforeSHA256":"7cab542474232d191a291fd42ab397e8a90cb7de1f204d1cbdcbb02807a5457f","afterSHA256":"7cab542474232d191a291fd42ab397e8a90cb7de1f204d1cbdcbb02807a5457f","reason":"Same source body in authored semantic page"}],"blocks":[{"id":"b-8e7e4a754c96272f85ed9ab1","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-8e7e4a754c96272f85ed9ab1","content":"[CLS 03 · 视觉与语言：任务、训练与接口](/course/notebooks/680da245145447dd88983aabd136fd24.html)","sha256":"343fb71807f077069b41c6495a8824ea4d353c740175b5775ccc5145b3bf19ca","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-4cafcedac2a9055b7b5a9009","prerequisites":[],"relatedBlocks":[]},{"id":"b-4cafcedac2a9055b7b5a9009","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-4cafcedac2a9055b7b5a9009","content":"# 学习路径： 文字候选与图文检索","sha256":"cf5bb768e672561b711388225dafb192390defe7ee096fd0c43eb07b12326246","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-8e7e4a754c96272f85ed9ab1","nextBlock":"b-663d90a41011b1e6eace0423","prerequisites":[{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-663d90a41011b1e6eace0423","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-663d90a41011b1e6eace0423","content":"主线阅读与优先实验","sha256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4cafcedac2a9055b7b5a9009","nextBlock":"b-bf239ab282cf8ee448fe420b","prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-bf239ab282cf8ee448fe420b","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-bf239ab282cf8ee448fe420b","content":"caption历史、VQA、CLIP候选、关系控制、GLIP/LLaVA与生成桥梁","sha256":"41c70c2c07b82678ff0a60f2d009b6afa750c5c691d109e754a69abf23868b50","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-663d90a41011b1e6eace0423","nextBlock":"b-3db83c6216a1320914e49251","prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-3db83c6216a1320914e49251","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-3db83c6216a1320914e49251","content":"**学习路径：** [文字候选与图文检索](/course/study/clip-retrieval.html)","sha256":"bab4bd804571aa5afbaf197e9c1539510fc766578c9825aa46cadbb18444cc32","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-bf239ab282cf8ee448fe420b","nextBlock":"b-7ad29d8c288e51bfb53bf83f","prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-7ad29d8c288e51bfb53bf83f","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-7ad29d8c288e51bfb53bf83f","content":"同一张图可以配文字、生成描述、回答问题，也可以按短语返回区域。本册把这四种输出放到同一条线上：语言此刻是答案、学习信号，还是任务条件？","sha256":"f727b99f9d6274f0111bb10a0383f86a10271b27674851de4ca88872fc35141a","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-3db83c6216a1320914e49251","nextBlock":"b-edac41e4bc5c95ad18a03ea9","prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-edac41e4bc5c95ad18a03ea9","type":"table","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-edac41e4bc5c95ad18a03ea9","content":"| 你想解决的问题 | 先读 | 再做或深读 |\n| --- | --- | --- |\n| 图像怎样改变一句话或一个答案 | [P51 描述](/course/pages/l3-r2-p51.html)、[P54 问答](/course/pages/l3-r2-p54.html) | [CLS 09](https://deepnote.com/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/c9f328a00ef44ecc90804e47276a7811)，按兴趣选 LAB13 或 LAB14 |\n| 图文配对怎样得到可复用表示 | [P57 矩阵](/course/pages/l3-r2-p57.html)、[P58 候选](/course/pages/l3-r2-p58.html) | 优先 [LAB 04](https://deepnote.com/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/0ba963931ddf4866b507b210d93a43bf)；需要图库缓存时再进 LAB12 |\n| 一句话指哪个对象，视觉如何接入语言模型 | [P63 定位](/course/pages/l3-r2-p63.html)、[P64 接口](/course/pages/l3-r2-p64.html) | [CLS 11](https://deepnote.com/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/95169c3e298c4c819373bb4534f3dff9)，其中 LAB15 是可复算的关系规则练习 |","sha256":"1d235867f8149bef480564be89cbe45135e146a1a5ca2cd3ac79ed658b91718b","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"table":{"headers":["你想解决的问题","先读","再做或深读"],"rows":[["你想解决的问题","先读","再做或深读"],["图像怎样改变一句话或一个答案","P51 描述、P54 问答","CLS 09，按兴趣选 LAB13 或 LAB14"],["图文配对怎样得到可复用表示","P57 矩阵、P58 候选","优先 LAB 04；需要图库缓存时再进 LAB12"],["一句话指哪个对象，视觉如何接入语言模型","P63 定位、P64 接口","CLS 11，其中 LAB15 是可复算的关系规则练习"]]},"previousBlock":"b-7ad29d8c288e51bfb53bf83f","nextBlock":"b-f15d6e975fbb4775abb90120","prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-f15d6e975fbb4775abb90120","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d.html#b-f15d6e975fbb4775abb90120","content":"读完应能为一个任务说清输入、训练目标、输出和适合的检验；不要求把所有分支顺序运行。本册的矩阵与规则代码是小算例，真实CLIP模型证据在LAB04。最后用[条件多解](/course/blocks/generation-one-condition-many-results.html)连接第四讲；用[研究划分小例](/course/blocks/research-question-split-evidence.html)检验自己的评价是否对应未来场景。","sha256":"7cab542474232d191a291fd42ab397e8a90cb7de1f204d1cbdcbb02807a5457f","chapter":{"id":"course","title":"学习路径： 文字候选与图文检索"},"unit":{"id":"notebook-blocks/009d0db9c9574995bdb7be7a82e66c8d","title":"学习路径： 文字候选与图文检索"},"page":{"id":"reading-opening","title":"学习路径： 文字候选与图文检索","number":1},"sectionTitle":"学习路径： 文字候选与图文检索","currentQuestion":"学习路径： 文字候选与图文检索","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-edac41e4bc5c95ad18a03ea9","nextBlock":null,"prerequisites":[{"id":"b-4cafcedac2a9055b7b5a9009","url":"#b-4cafcedac2a9055b7b5a9009","role":"section-title"},{"id":"b-8e7e4a754c96272f85ed9ab1","url":"#b-8e7e4a754c96272f85ed9ab1","role":"problem-context"}],"relatedBlocks":[]}]}