代寫(xiě)AIML 2023-2024 Coursework
時(shí)間：2024-03-17 來(lái)源： 作者： 我要糾錯(cuò)
AIML 2023-2024 Coursework
March 12, 2024
Figure 1: Convolutional neural network for coursework assignment.
Problem The goal of this take-home assignment is to implement, in Python, a simple two-layer convolutional neural network (CNN) with five inputs x1, . . . , x5, four hidden nodes z1, . . . , z4 and one output y with ReLU activations, according to the diagram shown in Figure 1. The hidden layer and output of the CNN is to be computed along with the gradient of the hidden layer and output with respect to parameter w1. The values oftheparameterswillbew1 =1.2,w2 =−0.2,v1 =−0.3,v2 =0.6,v3 =1.3andv4 =−1.5.
Instructions The CNN implementation is to be computed using a single Python function in single Python file. The interface to the function should be in the precise format,
y, z = convnet(x) (1)
where x = [x1, x2, x3, x4, x5] is a list of five numerical inputs (for example, a set of real numbers x=[0.3,−1.5,0.7,2.1,0.1]), and it should return the value of y as a number of the type dual and, z=[z1,z2,z3,z4] as a list of four numbers of type dual defined in the course code module ad.py. Therefore, when testing, you should expect to import this module. The implementation should use the specific values of the weight parameters given above.
Submission TopreparethePythoncodefileforsubmission,itmustbenamedintheformatinitials_studentid.py, for instance if your initials are ’AJD’ and your ID is 5716631 then your file should be named ajd_5716631.py. Submit the file through the Assignments page on Canvas. The deadline for submissions is 12pm UK time, 21st March 2024.
Marking The function will be marked automatically by calling it inside Python, and checking the results against a model solution. A fully correct solution will receive 20 marks. A solution which has a partially correct
請(qǐng)加QQ：99515681 郵箱：99515681@qq.com WX：codehelp
March 12, 2024
Figure 1: Convolutional neural network for coursework assignment.
Problem The goal of this take-home assignment is to implement, in Python, a simple two-layer convolutional neural network (CNN) with five inputs x1, . . . , x5, four hidden nodes z1, . . . , z4 and one output y with ReLU activations, according to the diagram shown in Figure 1. The hidden layer and output of the CNN is to be computed along with the gradient of the hidden layer and output with respect to parameter w1. The values oftheparameterswillbew1 =1.2,w2 =−0.2,v1 =−0.3,v2 =0.6,v3 =1.3andv4 =−1.5.
Instructions The CNN implementation is to be computed using a single Python function in single Python file. The interface to the function should be in the precise format,
y, z = convnet(x) (1)
where x = [x1, x2, x3, x4, x5] is a list of five numerical inputs (for example, a set of real numbers x=[0.3,−1.5,0.7,2.1,0.1]), and it should return the value of y as a number of the type dual and, z=[z1,z2,z3,z4] as a list of four numbers of type dual defined in the course code module ad.py. Therefore, when testing, you should expect to import this module. The implementation should use the specific values of the weight parameters given above.
Submission TopreparethePythoncodefileforsubmission,itmustbenamedintheformatinitials_studentid.py, for instance if your initials are ’AJD’ and your ID is 5716631 then your file should be named ajd_5716631.py. Submit the file through the Assignments page on Canvas. The deadline for submissions is 12pm UK time, 21st March 2024.
Marking The function will be marked automatically by calling it inside Python, and checking the results against a model solution. A fully correct solution will receive 20 marks. A solution which has a partially correct
請(qǐng)加QQ：99515681 郵箱：99515681@qq.com WX：codehelp
標(biāo)簽：
掃一掃在手機(jī)打開(kāi)當(dāng)前頁(yè)
注：本網(wǎng)條致力提供真實(shí)有用信息，所轉(zhuǎn)載的內(nèi)容，其版權(quán)均由原作者和資料提供方所擁有！若有任何不適煩請(qǐng)聯(lián)系我們，將會(huì)在24小時(shí)內(nèi)刪除。
無(wú)相關(guān)信息 昆明生活資訊
昆明圖文信息
|
|
|
|
|
|
|
|
推薦信息
相關(guān)文章
無(wú)相關(guān)信息
欄目更新