摘要: 改进模型后准确率变0.2623了 原本预训练模型训练6代草率了 第二次训练时候提高1迭代次数修改为200次 CUDA_ava=0,1,2,3 python main.py --n_epochs 200 \ --exp_name edm_competition \ --n_stability_samp 阅读全文
posted @ 2025-04-24 00:16 屈臣 阅读(10) 评论(0) 推荐(0)
摘要: 卷积层和池化层的实现 四维数组 from collections import OrderedDict import numpy as np from torch import nn x = np.random.rand(10,1,27,28)#随机生成数据 # x.shape x[0,0] 基于i 阅读全文
posted @ 2025-03-06 21:15 屈臣 阅读(12) 评论(0) 推荐(0)
摘要: 随机梯度下降的实现 class SGD: def __init__(self, lr=0.01): self.lr = lr def update(self,params,grads): for key in params.keys(): params[key] -= self.lr * grads 阅读全文
posted @ 2025-03-06 21:12 屈臣 阅读(2) 评论(0) 推荐(0)
摘要: 乘法层 class MulLayer: def __init__(self): self.x = None self.y = None def forward(self,x,y): self.x = x self.y = y out = x*y return out def backward(sel 阅读全文
posted @ 2025-03-06 21:05 屈臣 阅读(5) 评论(0) 推荐(0)
摘要: 一个二元函数的图像 import numpy as np import matplotlib.pyplot as plt # 生成数据 x0 = np.linspace(-3, 3, 100) x1 = np.linspace(-3, 3, 100) X0, X1 = np.meshgrid(x0, 阅读全文
posted @ 2025-03-06 21:00 屈臣 阅读(7) 评论(0) 推荐(0)
摘要: y = [0.1,0.05,0.6,0.0,0.05,0.1,0.0,0.1,0.0,0.0] t = [0,0,1,0,0,0,0,0,0,0] 均方误差 import numpy as np def mean_squared_error(y,t): return 0.5*np.sum((y-t) 阅读全文
posted @ 2025-03-06 20:53 屈臣 阅读(6) 评论(0) 推荐(0)
摘要: 基于pytorch的框架实现 导入必要的库 import torch import torchvision from torch import optim from torchvision import datasets, transforms from torch.utils.data impor 阅读全文
posted @ 2025-03-06 17:05 屈臣 阅读(20) 评论(0) 推荐(0)
摘要: softmax函数实现 import numpy as np a = np.array([0.3,2.9,4.0]) exp_a = np.exp(a) print(exp_a) sum_exp_a = np.sum(exp_a) print(sum_exp_a) y = exp_a / sum_e 阅读全文
posted @ 2025-03-06 16:46 屈臣 阅读(28) 评论(0) 推荐(0)
摘要: 须知:使用jupyter notebook食用 阶跃函数 def step_function(x): if x > 0: return 1 else: return 0 转换成支持Numpy数组的实现 import numpy as np def step_function_2(x): y = x> 阅读全文
posted @ 2025-03-04 20:53 屈臣 阅读(13) 评论(0) 推荐(0)
摘要: 与门 def AND(x1,x2): w1,w2,theta = 0.5,0.5,0.7 tmp = w1*x1+w2*x2 if tmp > theta: return 1 else: return 0 print(AND(0,1)) print(AND(1,0)) print(AND(1,1)) 阅读全文
posted @ 2025-03-04 20:30 屈臣 阅读(11) 评论(0) 推荐(0)