{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html","markdown":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.md","pages":[{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","unit":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","unitId":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","chapter":"course","number":1}],"aliases":[{"oldId":"b-014e3a8cd393549ff336971b","targetId":"b-014e3a8cd393549ff336971b","page":"reading-opening","operation":"move","beforeSHA256":"343aaa499ee36d5531eab9b8b2e10a88939c1d9efb4bb1251925f2c15e332b89","afterSHA256":"343aaa499ee36d5531eab9b8b2e10a88939c1d9efb4bb1251925f2c15e332b89","reason":"Same source body in authored semantic page"},{"oldId":"b-2d25848c2ee36ad0f4b90cd6","targetId":"b-2d25848c2ee36ad0f4b90cd6","page":"reading-opening","operation":"move","beforeSHA256":"7c642b153e037b8955b055ffd2cf3f9ed52e8402ed1a2da9927c7ae44185cdb4","afterSHA256":"7c642b153e037b8955b055ffd2cf3f9ed52e8402ed1a2da9927c7ae44185cdb4","reason":"Same source body in authored semantic page"},{"oldId":"b-26bd8cab749faa4778c9493c","targetId":"b-26bd8cab749faa4778c9493c","page":"reading-opening","operation":"move","beforeSHA256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","afterSHA256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","reason":"Same source body in authored semantic page"},{"oldId":"b-40afdb49d56a65bd975425ed","targetId":"b-40afdb49d56a65bd975425ed","page":"reading-opening","operation":"move","beforeSHA256":"e5185ec67dda4a18681db2d55dc5cbe0c312c9a5519949e3ded12e96d49c997f","afterSHA256":"e5185ec67dda4a18681db2d55dc5cbe0c312c9a5519949e3ded12e96d49c997f","reason":"Same source body in authored semantic page"},{"oldId":"b-95b87353a12168d34dac3421","targetId":"b-95b87353a12168d34dac3421","page":"reading-opening","operation":"move","beforeSHA256":"3223f363aa57cc36ff4abcc73f1079d3ca67df50ebe65daa792c552db83364c6","afterSHA256":"3223f363aa57cc36ff4abcc73f1079d3ca67df50ebe65daa792c552db83364c6","reason":"Same source body in authored semantic page"},{"oldId":"b-fb776706943da3ac61d89ada","targetId":"b-fb776706943da3ac61d89ada","page":"reading-opening","operation":"move","beforeSHA256":"035f14a00a62315b81cea78b634cb5234d1ce7590580c9c03727a7ee3431c706","afterSHA256":"035f14a00a62315b81cea78b634cb5234d1ce7590580c9c03727a7ee3431c706","reason":"Same source body in authored semantic page"},{"oldId":"b-6da5266fb1d29e48588168a0","targetId":"b-6da5266fb1d29e48588168a0","page":"reading-opening","operation":"move","beforeSHA256":"640df134a9e15a2ee3b6e15adfe56b56216f0113065cb39db8de3cff1ffae5b6","afterSHA256":"640df134a9e15a2ee3b6e15adfe56b56216f0113065cb39db8de3cff1ffae5b6","reason":"Same source body in authored semantic page"},{"oldId":"b-2097e8eb38a9c3eff20c1e0a","targetId":"b-2097e8eb38a9c3eff20c1e0a","page":"reading-opening","operation":"move","beforeSHA256":"7bbe824ac04c9ea2f884e59f0dd64895ba6168fb7bfd9ca6faafa40f9d8347ff","afterSHA256":"7bbe824ac04c9ea2f884e59f0dd64895ba6168fb7bfd9ca6faafa40f9d8347ff","reason":"Same source body in authored semantic page"}],"blocks":[{"id":"b-014e3a8cd393549ff336971b","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-014e3a8cd393549ff336971b","content":"[CLS 01 · 学习信号、表示与迁移](/course/notebooks/fb4365d604974dd19df492cb22516e44.html)","sha256":"343aaa499ee36d5531eab9b8b2e10a88939c1d9efb4bb1251925f2c15e332b89","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-2d25848c2ee36ad0f4b90cd6","prerequisites":[],"relatedBlocks":[]},{"id":"b-2d25848c2ee36ad0f4b90cd6","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-2d25848c2ee36ad0f4b90cd6","content":"# 我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sha256":"7c642b153e037b8955b055ffd2cf3f9ed52e8402ed1a2da9927c7ae44185cdb4","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-014e3a8cd393549ff336971b","nextBlock":"b-26bd8cab749faa4778c9493c","prerequisites":[{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-26bd8cab749faa4778c9493c","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-26bd8cab749faa4778c9493c","content":"主线阅读与优先实验","sha256":"42cba31deccba2a350b9b23bd0ca1dcaa4c4c5cb9d35c773e1fcb89124788ccd","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-2d25848c2ee36ad0f4b90cd6","nextBlock":"b-40afdb49d56a65bd975425ed","prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-40afdb49d56a65bd975425ed","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-40afdb49d56a65bd975425ed","content":"标签供给、文本目标接口、视图/队列/MAE图与表示迁移证据","sha256":"e5185ec67dda4a18681db2d55dc5cbe0c312c9a5519949e3ded12e96d49c997f","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-26bd8cab749faa4778c9493c","nextBlock":"b-95b87353a12168d34dac3421","prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-95b87353a12168d34dac3421","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-95b87353a12168d34dac3421","content":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建。计算组织使训练可实施，用途检验表示价值。","sha256":"3223f363aa57cc36ff4abcc73f1079d3ca67df50ebe65daa792c552db83364c6","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-40afdb49d56a65bd975425ed","nextBlock":"b-fb776706943da3ac61d89ada","prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-fb776706943da3ac61d89ada","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-fb776706943da3ac61d89ada","content":"现在固定一份可用的视觉表示，改变要求：**不只判断画面里有什么，还要找到每个对象的位置与边界。** 历史上的 R-CNN 从有监督分类预训练出发，再用检测数据适配；检测并不以自监督为前提。","sha256":"035f14a00a62315b81cea78b634cb5234d1ce7590580c9c03727a7ee3431c706","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-95b87353a12168d34dac3421","nextBlock":"b-6da5266fb1d29e48588168a0","prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-6da5266fb1d29e48588168a0","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-6da5266fb1d29e48588168a0","content":"接下来追问：哪些计算可以复用，哪些空间信息必须留下，新的输出需要怎样的训练？","sha256":"640df134a9e15a2ee3b6e15adfe56b56216f0113065cb39db8de3cff1ffae5b6","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-fb776706943da3ac61d89ada","nextBlock":"b-2097e8eb38a9c3eff20c1e0a","prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-2097e8eb38a9c3eff20c1e0a","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/d206f668feb646b2af65c349f020a9ba.html#b-2097e8eb38a9c3eff20c1e0a","content":"**选择一个继续入口：** 到 L03-CLS 02 看同一桌面场景的四种输出；到 L03-LAB 01 / 02 阅读已有真实结果；对无标签学习的前史感兴趣，可打开 LAB 05，看 Lloyd 如何实际拟合 patch 词典。","sha256":"7bbe824ac04c9ea2f884e59f0dd64895ba6168fb7bfd9ca6faafa40f9d8347ff","chapter":{"id":"course","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"unit":{"id":"notebook-blocks/d206f668feb646b2af65c349f020a9ba","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建"},"page":{"id":"reading-opening","title":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","number":1},"sectionTitle":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","currentQuestion":"我们已经看到几种训练表示的方法：人工标签、从文本构造目标、图像视图关系、遮挡重建","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-6da5266fb1d29e48588168a0","nextBlock":null,"prerequisites":[{"id":"b-2d25848c2ee36ad0f4b90cd6","url":"#b-2d25848c2ee36ad0f4b90cd6","role":"section-title"},{"id":"b-014e3a8cd393549ff336971b","url":"#b-014e3a8cd393549ff336971b","role":"problem-context"}],"relatedBlocks":[]}]}