fQwQf pushed 2 commits to master fQwQf/faQnet

October 24, 2024 6:18am

• 继续debug 34d00dc
• debug,debug! e4dcd31

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 23, 2024 3:43pm

• debug，但是没改完，剩下的明天再说 22920ca

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 23, 2024 3:24pm

• 重构神经网络反向传播算法 c247b92

View on GitHub

fQwQf pushed 4 commits to master fQwQf/faQnet

October 23, 2024 3:10pm

• 从底层重构神经网络 fed0a48
• 重构各个方法 f29499b
• 重构神经网络单层反向传播 3ef7cc6
• 重构神经网络梯度下降算法 b3081ee

View on GitHub

fQwQf pushed 2 commits to master fQwQf/faQnet

October 23, 2024 2:04pm

• 实现损失函数导数 54fb5bc
• 先声明 b6dcff0

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 23, 2024 1:29pm

• 之前是怎么跑通的？ b7baa3f

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 23, 2024 1:04pm

• 文件结构小调整 2e1754e

View on GitHub

fQwQf pushed 7 commits to master fQwQf/faQnet

October 23, 2024 12:59pm

• 实现softmax函数及其导数计算 cb0addd
• 优化了平均绝对误差函数的实现 c2bc681
• 优化了平均平方误差函数的实现 383c790
• 优化了平均绝对百分比误差函数的实现 eff0384
• 优化了平滑平均绝对误差函数的实现 6a68021
• 优化了平均平方误差函数的实现 b409948
• 优化了二元交叉熵损失函数的实现 4aee8d7

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 4:57pm

• 大幅优化反向传播算法 297566e

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 4:41pm

• 增加了损失函数的导函数 79c3840

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:48pm

• 日常开发 1fabe2f

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:32pm

• 添加了结果可视化的构想 5c3e1ed

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:27pm

• 修改措辞 3e0e942

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:22pm

• 提高对归一化方法的预期 6058547

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:17pm

• 已经实现的就不用放在todolist里了 ac2e410

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:12pm

• 优化了反向传播的误差计算 e7b0246

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 3:07pm

• 优化最后一层的运算 e6ba522

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 22, 2024 1:50pm

• 提交一下，证明我还活着 2941a99

View on GitHub

stephen-zeng starred fQwQf/faQnet

October 18, 2024 11:16am

fQwQf pushed 2 commits to master fQwQf/faQnet

October 18, 2024 2:49am

• 添加了softmax层 5810565
• 总不能把训练集当测试集吧 582e1f2

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 17, 2024 5:32am

• 实现了softmax函数 9c7a345

View on GitHub

fQwQf pushed 1 commit to master fQwQf/faQnet

October 17, 2024 5:19am

• 日常开发 4f60144

View on GitHub

fQwQf starred fQwQf/faQnet

October 16, 2024 3:21pm

fQwQf pushed 2 commits to master fQwQf/faQnet

October 16, 2024 3:02pm

• loss训练函数添加了损失阈值参数 f3925f7
• 把数据也提交了算了 1f4b365

View on GitHub

fQwQf pushed 3 commits to master fQwQf/faQnet

October 16, 2024 12:24pm

• 调整了损失函数 4e239e4
• 实现了按loss值训练和整个神经网络初始化 72a185a
• 修复了损失函数打印 34cc346

View on GitHub

fQwQf pushed 2 commits to master fQwQf/faQnet

October 15, 2024 3:53pm

• 添加了新的图片文件 2451df3
• 日常开发 6e86a67

View on GitHub

fQwQf pushed 2 commits to master fQwQf/faQnet

October 15, 2024 3:47pm

• 修复了SLL和MSLE函数 f6499e0
• 修复了CE函数，这也意味着所有函数修复完成 fab4b9e

View on GitHub