使用多个文件训练机器学习模型

发布于 2025-01-18 04:22:36 字数 425 浏览 1 评论 0原文

我是深度学习方面的新手，特别是每次处理多个文件。我有一个文件文件夹，我希望我的模型（LSTM）文件使用所有这些文件进行训练。

我可以在路径文件的循环中使用 model.fit() 吗？这有意义吗？或者我应该尝试将其连接到一个数组中？在这种情况下，问题是特征具有不同的范围。

import os 
rootdir = '/content/Data' #looping over files
for filename in os.listdir('/content/Data'):
 with open(os.path.join(rootdir, filename)) as f:   
  A = np.loadtxt(f)     model.fit(x_train, y_train, batch_size=50, 
  epochs=20, shuffle=True)

原文

I'm kind of new in working with deep learning, and specially with more than one file per time. I have a folder of files, and I want that my model(LSTM) files train with all that files.

Can I just use model.fit() inside a loop of a path files? Would it make sense? Or I should try to concatenate that into a single array? In this case the problem is the features have different ranges.

import os 
rootdir = '/content/Data' #looping over files
for filename in os.listdir('/content/Data'):
 with open(os.path.join(rootdir, filename)) as f:   
  A = np.loadtxt(f)     model.fit(x_train, y_train, batch_size=50, 
  epochs=20, shuffle=True)

分享到QQ

分享到微博