{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"已完成结果：复杂模型没有被安排成赢家","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html","markdown":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.md","pages":[{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","unit":"已完成结果：复杂模型没有被安排成赢家","unitId":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","chapter":"course","number":1}],"aliases":[{"oldId":"b-f5f7c90853c98a3443f6104f","targetId":"b-f5f7c90853c98a3443f6104f","page":"reading-opening","operation":"move","beforeSHA256":"f6dfd09a50caab375f02e0abcc3ff7a454ea579ed8eef0c35b95c95b0619c304","afterSHA256":"f6dfd09a50caab375f02e0abcc3ff7a454ea579ed8eef0c35b95c95b0619c304","reason":"Same source body in authored semantic page"},{"oldId":"b-1511df36b9cdae713a104993","targetId":"b-1511df36b9cdae713a104993","page":"reading-opening","operation":"move","beforeSHA256":"6f30239b136053845c3c7a3a6bcf186cc10d994441f4ce2699c0c4ef1e6cd198","afterSHA256":"6f30239b136053845c3c7a3a6bcf186cc10d994441f4ce2699c0c4ef1e6cd198","reason":"Same source body in authored semantic page"},{"oldId":"b-dfbe7c9ecd38cf015aff3b11","targetId":"b-dfbe7c9ecd38cf015aff3b11","page":"reading-opening","operation":"move","beforeSHA256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","afterSHA256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","reason":"Same source body in authored semantic page"},{"oldId":"b-a98fcda1fd73985adedef3c7","targetId":"b-a98fcda1fd73985adedef3c7","page":"reading-opening","operation":"move","beforeSHA256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","afterSHA256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","reason":"Same source body in authored semantic page"},{"oldId":"b-17dc3a8fd12f4d51052ba280","targetId":"b-17dc3a8fd12f4d51052ba280","page":"reading-opening","operation":"move","beforeSHA256":"16424645f3255bd542eaba92027ec7ebc3e44924856df20f0677d9042a3a9af4","afterSHA256":"16424645f3255bd542eaba92027ec7ebc3e44924856df20f0677d9042a3a9af4","reason":"Same source body in authored semantic page"},{"oldId":"b-eb08f34960fee9c2f6a94a5b","targetId":"b-eb08f34960fee9c2f6a94a5b","page":"reading-opening","operation":"move","beforeSHA256":"03a19cae81d2ff11ebeebadbfcc0fe5635e45af05769809a745eaab573b92218","afterSHA256":"03a19cae81d2ff11ebeebadbfcc0fe5635e45af05769809a745eaab573b92218","reason":"Same source body in authored semantic page"},{"oldId":"b-d7ae2bf1b38fdc76a6729fcb","targetId":"b-d7ae2bf1b38fdc76a6729fcb","page":"reading-opening","operation":"rewrite","beforeSHA256":"0476bb05277a04c23e4c8b6ce52d39f57fa7c289952f4eb13dfa979376725eae","afterSHA256":"c1b378d36ab88faa10dd2f0b6326a5e65b6738a79d88081bf4be163e38fec080","reason":"Same source identity with figure semantics or reading context enriched"},{"oldId":"b-2cf492f204f5dd3e8f7da5fc","targetId":"b-2cf492f204f5dd3e8f7da5fc","page":"reading-opening","operation":"move","beforeSHA256":"101e0c7e5c59f819cddc0f9d349f0e531473f5f092cf9b4dff94298693faa83f","afterSHA256":"101e0c7e5c59f819cddc0f9d349f0e531473f5f092cf9b4dff94298693faa83f","reason":"Same source body in authored semantic page"},{"oldId":"b-0a6bf83eafc08ed5ed780582","targetId":"b-0a6bf83eafc08ed5ed780582","page":"reading-opening","operation":"move","beforeSHA256":"e465cfb291fd252309d7f2181ce240583761f7b6d30f4234070d6fb5e87896ff","afterSHA256":"805f9aef538f2fbdbb2a98574452c74ecb9fae28fbe7bff0adbce451952c0d01","reason":"Same source body in authored semantic page"},{"oldId":"b-352706456fabb157ecf84d70","targetId":"b-352706456fabb157ecf84d70","page":"reading-opening","operation":"move","beforeSHA256":"f51eac1aae8965802b84880530a9270d8c3e3751e42dbde5940e8d401b8c16f2","afterSHA256":"cab5afe03f707fb5bbca4cddf8778f536cfb8d6b9adca57c29a5107c47636a95","reason":"Same source body in authored semantic page"},{"oldId":"b-f34017f3da715744dfd69277","targetId":"b-f34017f3da715744dfd69277","page":"reading-opening","operation":"move","beforeSHA256":"26ff18341d4fa15f7e4403a265645c5904a90634a1aaa35f64c7506e131fb06d","afterSHA256":"26ff18341d4fa15f7e4403a265645c5904a90634a1aaa35f64c7506e131fb06d","reason":"Same source body in authored semantic page"},{"oldId":"b-37d81ce9cfd58eca708cdf0e","targetId":"b-37d81ce9cfd58eca708cdf0e","page":"reading-opening","operation":"move","beforeSHA256":"334be2c197ec023e6708646506157991473e880d4347d6e99d48afd83448f2f9","afterSHA256":"cdaa1a9b8ed2f648a296b64726de36de5bb79d8decf910109b5bb708b8966142","reason":"Same source body in authored semantic page"},{"oldId":"b-ddfd6d21cc58e4ff4e9224ec","targetId":"b-ddfd6d21cc58e4ff4e9224ec","page":"reading-opening","operation":"move","beforeSHA256":"9412c9cf2be0c73ea8850a70b998b42219a9f8baf916c61ba9a7c9d760da4526","afterSHA256":"9412c9cf2be0c73ea8850a70b998b42219a9f8baf916c61ba9a7c9d760da4526","reason":"Same source body in authored semantic page"},{"oldId":"b-f43f0cc3130dfdd7bb6d6b27","targetId":"b-f43f0cc3130dfdd7bb6d6b27","page":"reading-opening","operation":"move","beforeSHA256":"cb5121b7a60ed7a947e6ba295f10c4ff937fcc81522c6c76df945b092efc2ba7","afterSHA256":"cef1fd6369aabb1d23e9ec0152605ca225c6913875d83c223eae49ad8214d1d2","reason":"Same source body in authored semantic page"},{"oldId":"b-240f8a01bcc4dfb5205e7ffc","targetId":"b-240f8a01bcc4dfb5205e7ffc","page":"reading-opening","operation":"move","beforeSHA256":"872af66a9cbfc66762f9bf89802b01cd9d1317d99427da3c4957ed638521532a","afterSHA256":"872af66a9cbfc66762f9bf89802b01cd9d1317d99427da3c4957ed638521532a","reason":"Same source body in authored semantic page"},{"oldId":"block-16412ad68e504e8f9b564ed686a53b4c--已完成结果复杂模型没有被安排成赢家","targetId":"b-17dc3a8fd12f4d51052ba280","operation":"retain-source-identity"}],"blocks":[{"id":"b-f5f7c90853c98a3443f6104f","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-f5f7c90853c98a3443f6104f","content":"[LAB 11 · 从 RGB 到像素：读出、FCN 与边界](/course/notebooks/9e268706471944e3a59e6f568f9ce9a3.html)","sha256":"f6dfd09a50caab375f02e0abcc3ff7a454ea579ed8eef0c35b95c95b0619c304","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-1511df36b9cdae713a104993","prerequisites":[],"relatedBlocks":[]},{"id":"b-1511df36b9cdae713a104993","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-1511df36b9cdae713a104993","content":"# 已完成结果：复杂模型没有被安排成赢家","sha256":"6f30239b136053845c3c7a3a6bcf186cc10d994441f4ce2699c0c4ef1e6cd198","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f5f7c90853c98a3443f6104f","nextBlock":"b-dfbe7c9ecd38cf015aff3b11","prerequisites":[{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-dfbe7c9ecd38cf015aff3b11","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-dfbe7c9ecd38cf015aff3b11","content":"按问题继续","sha256":"e91f24e4e4ca40e2099a0d81d30db6568f4bf2074b499eeee6e3959617bd97ac","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-1511df36b9cdae713a104993","nextBlock":"b-a98fcda1fd73985adedef3c7","prerequisites":[{"id":"b-1511df36b9cdae713a104993","url":"#b-1511df36b9cdae713a104993","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-a98fcda1fd73985adedef3c7","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-a98fcda1fd73985adedef3c7","content":"前景/背景/mIoU口径、RGB到mask训练、FCN与1×1读出及边界","sha256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-dfbe7c9ecd38cf015aff3b11","nextBlock":"b-17dc3a8fd12f4d51052ba280","prerequisites":[{"id":"b-1511df36b9cdae713a104993","url":"#b-1511df36b9cdae713a104993","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-17dc3a8fd12f4d51052ba280","type":"h2","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-17dc3a8fd12f4d51052ba280","content":"## 已完成结果：复杂模型没有被安排成赢家","sha256":"16424645f3255bd542eaba92027ec7ebc3e44924856df20f0677d9042a3a9af4","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-a98fcda1fd73985adedef3c7","nextBlock":"b-eb08f34960fee9c2f6a94a5b","prerequisites":[{"id":"b-1511df36b9cdae713a104993","url":"#b-1511df36b9cdae713a104993","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-eb08f34960fee9c2f6a94a5b","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-eb08f34960fee9c2f6a94a5b","content":"本地 CPU 完整运行开始于 `2026-09-22T16:28:24.497499+00:00`，PyTorch `2.14.0+cpu`；源码 SHA-256 为 `ecbe38c4a53eb691b7e31be95eaa8ab3a0b93578005d2605072196cc5dc66445`。FCN 训练 CE 从 .85632 降到 .00492。下面 mIoU 均对背景、红矩形、蓝椭圆三个类别等权平均；先在整份保留集累计混淆矩阵，再计算逐类 IoU，未对每张图的 IoU 简单平均。","sha256":"03a19cae81d2ff11ebeebadbfcc0fe5635e45af05769809a745eaab573b92218","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-17dc3a8fd12f4d51052ba280","nextBlock":"b-d7ae2bf1b38fdc76a6729fcb","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d7ae2bf1b38fdc76a6729fcb","type":"table","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-d7ae2bf1b38fdc76a6729fcb","content":"| 模型 / 输入条件 | 三类 mIoU | 结论能说到哪里 |\n| --- | --- | --- |\n| 随机初始化 FCN | .01470 | 尚未学到类别读出 |\n| 训练后 1×1 RGB 读出 | 1.00000 | 本合成集每个像素颜色已足以分类 |\n| 训练后 FCN，双线性 | .99041 | 学到图像到 mask，边界仍有少量错 |\n| 同 FCN，最近邻 logits | .91180 | 改变上采样坐标权重影响空间决策 |\n| 同 FCN，对象尺寸 ×.45 | .94396 | 细小蓝对象和窄缝更难保持 |\n| 同 FCN，RGB 5×5 模糊 | .92397 | 边缘观测变弱时性能下降 |\n| 同 RGB 读出，5×5 模糊 | .84996 | 无邻域基准受混合颜色更大影响 |","sha256":"c1b378d36ab88faa10dd2f0b6326a5e65b6738a79d88081bf4be163e38fec080","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"table":{"headers":["模型 / 输入条件","三类 mIoU","结论能说到哪里"],"rows":[["模型 / 输入条件","三类 mIoU","结论能说到哪里"],["随机初始化 FCN",".01470","尚未学到类别读出"],["训练后 1×1 RGB 读出","1.00000","本合成集每个像素颜色已足以分类"],["训练后 FCN，双线性",".99041","学到图像到 mask，边界仍有少量错"],["同 FCN，最近邻 logits",".91180","改变上采样坐标权重影响空间决策"],["同 FCN，对象尺寸 ×.45",".94396","细小蓝对象和窄缝更难保持"],["同 FCN，RGB 5×5 模糊",".92397","边缘观测变弱时性能下降"],["同 RGB 读出，5×5 模糊",".84996","无邻域基准受混合颜色更大影响"]]},"previousBlock":"b-eb08f34960fee9c2f6a94a5b","nextBlock":"b-2cf492f204f5dd3e8f7da5fc","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-2cf492f204f5dd3e8f7da5fc","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-2cf492f204f5dd3e8f7da5fc","content":"像素读出比 FCN 的原分布结果更好，是有意义的负结果：更复杂结构不是得分保证，数据又带有强颜色线索。在尺寸 ×.45 条件下读出仍为 1.0，说明纯颜色规则不会因为面积变小而失效；在模糊条件下，邻域 FCN 才表现出比单像素读出更强的耐受。这不能直接推广到真实语义分割，真实同类物体的颜色并不恒定。","sha256":"101e0c7e5c59f819cddc0f9d349f0e531473f5f092cf9b4dff94298693faa83f","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-d7ae2bf1b38fdc76a6729fcb","nextBlock":"b-0a6bf83eafc08ed5ed780582","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-0a6bf83eafc08ed5ed780582","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-0a6bf83eafc08ed5ed780582","content":"![采样热身](/assets/display/0eb14c8f3f86cf2d8b3c9b65.webp)","sha256":"805f9aef538f2fbdbb2a98574452c74ecb9fae28fbe7bff0adbce451952c0d01","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-2cf492f204f5dd3e8f7da5fc","nextBlock":"b-352706456fabb157ecf84d70","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-352706456fabb157ecf84d70","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-352706456fabb157ecf84d70","content":"![保留集逐像素预测](/assets/display/1b11f25259c1fe3405e0b7a5.webp)","sha256":"cab5afe03f707fb5bbca4cddf8778f536cfb8d6b9adca57c29a5107c47636a95","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-0a6bf83eafc08ed5ed780582","nextBlock":"b-f34017f3da715744dfd69277","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-f34017f3da715744dfd69277","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-f34017f3da715744dfd69277","content":"错误最多的图 89 有 25 个有效像素错，图 125 有 18 个，图 56 有 16 个，图 92 有 13 个。图 89/125 的两个蓝椭圆间出现背景窄缝被填平的错误：mask 所需的边界证据可能在低分辨率表示中混合，即使细分支能改善，也不保证完全恢复。此处蓝色连通不等于实例计数正确，本册没有训练实例分割。","sha256":"26ff18341d4fa15f7e4403a265645c5904a90634a1aaa35f64c7506e131fb06d","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-352706456fabb157ecf84d70","nextBlock":"b-37d81ce9cfd58eca708cdf0e","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-37d81ce9cfd58eca708cdf0e","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-37d81ce9cfd58eca708cdf0e","content":"![错误像素与窄缝](/assets/display/f2028b85a049b53efbcbbb7a.webp)","sha256":"cdaa1a9b8ed2f648a296b64726de36de5bb79d8decf910109b5bb708b8966142","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f34017f3da715744dfd69277","nextBlock":"b-ddfd6d21cc58e4ff4e9224ec","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-ddfd6d21cc58e4ff4e9224ec","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-ddfd6d21cc58e4ff4e9224ec","content":"尺寸 ×.45 后像素准确率仍有 .99816，但蓝椭圆 IoU 只有 .85763，红矩形 .97613，背景 .99812。小对象在整图像素数中占比太低，仅报告准确率会掩盖其损失。最近邻对照的蓝类 IoU 进一步降到 .80928，而背景 .98558，展示了不同类别边界受采样影响不同。","sha256":"9412c9cf2be0c73ea8850a70b998b42219a9f8baf916c61ba9a7c9d760da4526","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-37d81ce9cfd58eca708cdf0e","nextBlock":"b-f43f0cc3130dfdd7bb6d6b27","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-f43f0cc3130dfdd7bb6d6b27","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-f43f0cc3130dfdd7bb6d6b27","content":"![小对象条件](/assets/display/a3b94d7df0f7bd4472d4f87d.webp)","sha256":"cef1fd6369aabb1d23e9ec0152605ca225c6913875d83c223eae49ad8214d1d2","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ddfd6d21cc58e4ff4e9224ec","nextBlock":"b-240f8a01bcc4dfb5205e7ffc","prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-240f8a01bcc4dfb5205e7ffc","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/notebook-blocks/16412ad68e504e8f9b564ed686a53b4c.html#b-240f8a01bcc4dfb5205e7ffc","content":"下一次只把训练增强加入模糊，保持场景、epoch 和网络不变，再看“干净图/模糊图”两列是否同时改善。若想检验颜色捷径，应重新定义颜色与类别的生成关系并同时调整训练与测试协议；不能把标签含义改变的分布转移误说成边界采样问题。","sha256":"872af66a9cbfc66762f9bf89802b01cd9d1317d99427da3c4957ed638521532a","chapter":{"id":"course","title":"已完成结果：复杂模型没有被安排成赢家"},"unit":{"id":"notebook-blocks/16412ad68e504e8f9b564ed686a53b4c","title":"已完成结果：复杂模型没有被安排成赢家"},"page":{"id":"reading-opening","title":"已完成结果：复杂模型没有被安排成赢家","number":1},"sectionTitle":"已完成结果：复杂模型没有被安排成赢家","currentQuestion":"已完成结果：复杂模型没有被安排成赢家","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f43f0cc3130dfdd7bb6d6b27","nextBlock":null,"prerequisites":[{"id":"b-17dc3a8fd12f4d51052ba280","url":"#b-17dc3a8fd12f4d51052ba280","role":"section-title"},{"id":"b-f5f7c90853c98a3443f6104f","url":"#b-f5f7c90853c98a3443f6104f","role":"problem-context"}],"relatedBlocks":[]}]}