JavaBigData的博客

df.ix[:0,0] 更改为 df.iloc[:0,1]

import torch t = torch.tensor([2, 3]) print(t) x=torch.randn(3,4) print(x) # x_gpu=x.cuda(0) x_gpu2=x.cuda() print(x_gpu2) x_cpu=x_gpu2.cpu() print(x_cpu) x_arr=x_cpu.cpu().numpy() print(x_gpu2) print(x_arr)

卸载重装 pip uninstall tensorboard pip install tensorboard

import numpy as np import tensorflow as tf x_data=np.random.rand(1000) y_data=x_data*10 +77 print(x_data) print(y_data) k=tf.Variable(0.,name='k') b=tf.Variable(0.) y=k*x_data+b loss = tf.reduce_mean(tf.square(y-y_data)) optimizer = tf.train.GradientDes

反编译修改容易出错，建议用以下方式修改 pip uninstall jupyter jupyter-client jupyter-console jupyter-core jupyterlab-pygments jupyterlab-widgets 删除script目录下面的文件 然后安装 pip install jupyter ...

raw_text=open('Winston_Churchil.txt').read() ???? raw_text=open('Winston_Churchil.txt','r',encoding='UTF-8').read()

pytorch运行出错：AttributeError: module ‘torch‘ has no attribute ‘ops‘ pip install torchvision==0.4.1 -i https://pypi.tuna.tsinghua.edu.cn/simple/

参考其他大佬，在 @torch.jit.script 这个注释前加注释 ：# @torch.jit.script，可以解决！

python安装库报权限问题时使用pip install torch --user ,会将模块安装自己的文件目录下

机器性能太弱，不能设置太多参数，梯度下降难 import numpy as np import random from PIL import Image import matplotlib.pyplot as plt import tensorflow as tf def code_cnn(x,y): #conv->relu->max_pool->conv->relu->max_pool->dropout # ->conv->relu-&g

参数太多，gpu内存不足，适当减少参数

数量要一样

import os import random import numpy as np from captcha.image import ImageCaptcha from random import randint from PIL import Image import matplotlib.pyplot as plt code_char_set = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd