Troubleshooting a deploy to heroku - RuntimeError: The reset parameter is False but there is no n_features_in_ attribute. Is this estimator fitted?
Troubleshooting a deploy to heroku - RuntimeError: The reset parameter is False but there is no n_features_in_ attribute. Is this estimator fitted?
您好 Stackeoverflow 社区。我正在为 class 开发一个简单的项目，涉及托管一个允许用户输入值并预测谁将获胜的烧瓶应用程序。好吧，在我们使用 python flask 的本地机器上运行正常，但在 heroku 上提出了挑战。难倒了！
代码不长也不复杂。也知道我们也有一些机会。
我们的 App.py 文件看起来像：
import flask
import tensorflow
import sklearn
import pandas as pd
import numpy as np
from tensorflow import keras
from pickle import load
from load import *
model = init()
scaler = load(open('model/scaler.pkl', 'rb'))
print("Awesome, your scaler has been loaded from disk! Cool beans!")
app = flask.Flask(__name__, template_folder='templates')
@app.route('/', methods=['GET', 'POST'])
def main():
global model
if flask.request.method == 'GET':
return(flask.render_template('main.html'))
if flask.request.method == 'POST':
R_Weight = flask.request.form['R_Weight']
R_Height = flask.request.form['R_Height']
R_Age = flask.request.form['R_Age']
B_Weight = flask.request.form['B_Weight']
B_Height = flask.request.form['B_Height']
B_Age = flask.request.form['B_Age']
RPrev = 2.130049
BPrev = 1.756650
BStreak = 0.643350
RStreak = 0.748768
input_variables = pd.DataFrame([[BPrev, BStreak, B_Age,B_Height,B_Weight,RPrev, RStreak, R_Age,R_Height,R_Weight]],
columns=['BPrev','BStreak','B_Age','B_Height','B_Weight','RPrev','RStreak','R_Age','R_Height','R_Weight'],
dtype=float)
input_scaled = scaler.transform(input_variables)
prediction = model.predict(input_scaled)[0][0]
if np.round(prediction) == 0:
prediction = "Blue"
else:
prediction = "Red"
if int(R_Weight) < 20:
prediction = "Invalid Weight"
elif int(B_Weight) < 20:
prediction = "Invalid Weight"
elif int(R_Height) < 20:
prediction = "Invalid Height"
elif int(B_Height) < 20:
prediction = "Invalid Height"
elif int(R_Age) < 18:
prediction = "Invalid Age"
elif int(B_Age) < 18:
prediction = "Invalid Age"
return flask.render_template('main.html',
original_input={'R_Weight' :R_Weight,
'R_Height' :R_Height,
'R_Age' :R_Age,
'B_Weight' :B_Weight,
'B_Height' :B_Height,
'B_Age' :B_Age
},
result=prediction
)
@app.route('/logisticregression')
def logreg():
return(flask.render_template('00_model_logisticsregression.html'))
@app.route('/gridsearch')
def grisea():
return(flask.render_template('00_model_gridsearch.html'))
@app.route('/deeplearning')
def deelea():
return(flask.render_template('01_training_model.html'))
@app.route('/viz1')
def viz1():
return(flask.render_template('00_data_visualization_a.html'))
@app.route('/viz2')
def viz2():
return(flask.render_template('00_data_visualization_b.html'))
@app.route('/viz3')
def viz3():
return(flask.render_template('00_data_visualization_c.html'))
@app.route('/trained')
def traine():
return(flask.render_template('02_notraining_model.html'))
@app.route('/debugging')
def debugg():
return(flask.render_template('03_debugger.html'))
if __name__ == '__main__':
app.run(debug=True)
我在 heroku 中看到的错误是：
2020-05-31T19:43:05.000000+00:00 app[api]: Build succeeded
2020-05-31T19:48:29.426002+00:00 app[web.1]: 10.81.161.51 - - [31/May/2020:19:48:29 +0000] "GET / HTTP/1.1" 200 4710 "https://dashboard.heroku.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.432446+00:00 heroku[router]: at=info method=GET path="/" host=bcs-final.herokuapp.com request_id=1df55828-886a-45d1-9605-aafd3414f9e3 fwd="71.143.151.112" dyno=web.1 connect=1ms service=56ms status=200 bytes=4872 protocol=https
2020-05-31T19:48:29.623294+00:00 app[web.1]: 10.81.161.51 - - [31/May/2020:19:48:29 +0000] "GET /reset.css HTTP/1.1" 404 232 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.623792+00:00 heroku[router]: at=info method=GET path="/reset.css" host=bcs-final.herokuapp.com request_id=40816718-e41b-44e9-a3d5-da5ff2030adb fwd="71.143.151.112" dyno=web.1 connect=1ms service=7ms status=404 bytes=400 protocol=https
2020-05-31T19:48:29.625560+00:00 app[web.1]: 10.7.244.15 - - [31/May/2020:19:48:29 +0000] "GET /style.css HTTP/1.1" 404 232 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.627392+00:00 heroku[router]: at=info method=GET path="/style.css" host=bcs-final.herokuapp.com request_id=bd01a28c-7d1e-4dc6-9a85-405b1bdbb8ea fwd="71.143.151.112" dyno=web.1 connect=1ms service=8ms status=404 bytes=400 protocol=https
2020-05-31T19:48:37.727173+00:00 app[web.1]: [2020-05-31 19:48:37,723] ERROR in app: Exception on / [POST]
2020-05-31T19:48:37.727185+00:00 app[web.1]: Traceback (most recent call last):
2020-05-31T19:48:37.727186+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 2447, in wsgi_app
2020-05-31T19:48:37.727186+00:00 app[web.1]: response = self.full_dispatch_request()
2020-05-31T19:48:37.727187+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1952, in full_dispatch_request
2020-05-31T19:48:37.727215+00:00 app[web.1]: rv = self.handle_user_exception(e)
2020-05-31T19:48:37.727216+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1821, in handle_user_exception
2020-05-31T19:48:37.727217+00:00 app[web.1]: reraise(exc_type, exc_value, tb)
2020-05-31T19:48:37.727218+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/_compat.py", line 39, in reraise
2020-05-31T19:48:37.727218+00:00 app[web.1]: raise value
2020-05-31T19:48:37.727219+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1950, in full_dispatch_request
2020-05-31T19:48:37.727219+00:00 app[web.1]: rv = self.dispatch_request()
2020-05-31T19:48:37.727220+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1936, in dispatch_request
2020-05-31T19:48:37.727220+00:00 app[web.1]: return self.view_functions[rule.endpoint](**req.view_args)
2020-05-31T19:48:37.727221+00:00 app[web.1]: File "/app/app.py", line 36, in main
2020-05-31T19:48:37.727221+00:00 app[web.1]: input_scaled = scaler.transform(input_variables)
2020-05-31T19:48:37.727222+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/preprocessing/_data.py", line 794, in transform
2020-05-31T19:48:37.727222+00:00 app[web.1]: force_all_finite='allow-nan')
2020-05-31T19:48:37.727222+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/base.py", line 436, in _validate_data
2020-05-31T19:48:37.727223+00:00 app[web.1]: self._check_n_features(X, reset=reset)
2020-05-31T19:48:37.727223+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/base.py", line 373, in _check_n_features
2020-05-31T19:48:37.727223+00:00 app[web.1]: "The reset parameter is False but there is no "
2020-05-31T19:48:37.727314+00:00 app[web.1]: RuntimeError: The reset parameter is False but there is no n_features_in_ attribute. Is this estimator fitted?
2020-05-31T19:48:37.728944+00:00 app[web.1]: 10.7.244.15 - - [31/May/2020:19:48:37 +0000] "POST / HTTP/1.1" 500 290 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:37.730539+00:00 heroku[router]: at=info method=POST path="/" host=bcs-final.herokuapp.com request_id=2ba8ae51-f913-48d0-9b45-78aef407c15e fwd="71.143.151.112" dyno=web.1 connect=1ms service=27ms status=500 bytes=470 protocol=https
查看代码，我无法确定运行在本地和在 heroku 上使用您的应用程序之间有什么区别。唯一合理的猜测是您的本地环境与 heroku 上的 运行 版本不同。
无论如何，如果我们关注 heroku 的版本告诉我们的内容，您的错误来自这一行：
# ...
input_scaled = scaler.transform(input_variables) # <--- here
# ...
错误基本上是说：估计器 - 实际上是你的缩放器（参见源代码 here） - 不适合。
消息本身被打印出来 here 并且说“n_features_in_ 属性缺失”的部分本质上意味着您的 input_variables 稀疏矩阵（在 scikit-learn 源代码中称为 X) 缺少属性 n_features_in_。该属性设置为拟合过程的一部分。
最重要的是，在 transform 之前调用 fit，这可能会有所帮助。所以像：
input_scaled = scaler.fit(input_variables).transform(input_variables)
您好 Stackeoverflow 社区。我正在为 class 开发一个简单的项目，涉及托管一个允许用户输入值并预测谁将获胜的烧瓶应用程序。好吧，在我们使用 python flask 的本地机器上运行正常，但在 heroku 上提出了挑战。难倒了！
代码不长也不复杂。也知道我们也有一些机会。
我们的 App.py 文件看起来像：
import flask
import tensorflow
import sklearn
import pandas as pd
import numpy as np
from tensorflow import keras
from pickle import load
from load import *
model = init()
scaler = load(open('model/scaler.pkl', 'rb'))
print("Awesome, your scaler has been loaded from disk! Cool beans!")
app = flask.Flask(__name__, template_folder='templates')
@app.route('/', methods=['GET', 'POST'])
def main():
global model
if flask.request.method == 'GET':
return(flask.render_template('main.html'))
if flask.request.method == 'POST':
R_Weight = flask.request.form['R_Weight']
R_Height = flask.request.form['R_Height']
R_Age = flask.request.form['R_Age']
B_Weight = flask.request.form['B_Weight']
B_Height = flask.request.form['B_Height']
B_Age = flask.request.form['B_Age']
RPrev = 2.130049
BPrev = 1.756650
BStreak = 0.643350
RStreak = 0.748768
input_variables = pd.DataFrame([[BPrev, BStreak, B_Age,B_Height,B_Weight,RPrev, RStreak, R_Age,R_Height,R_Weight]],
columns=['BPrev','BStreak','B_Age','B_Height','B_Weight','RPrev','RStreak','R_Age','R_Height','R_Weight'],
dtype=float)
input_scaled = scaler.transform(input_variables)
prediction = model.predict(input_scaled)[0][0]
if np.round(prediction) == 0:
prediction = "Blue"
else:
prediction = "Red"
if int(R_Weight) < 20:
prediction = "Invalid Weight"
elif int(B_Weight) < 20:
prediction = "Invalid Weight"
elif int(R_Height) < 20:
prediction = "Invalid Height"
elif int(B_Height) < 20:
prediction = "Invalid Height"
elif int(R_Age) < 18:
prediction = "Invalid Age"
elif int(B_Age) < 18:
prediction = "Invalid Age"
return flask.render_template('main.html',
original_input={'R_Weight' :R_Weight,
'R_Height' :R_Height,
'R_Age' :R_Age,
'B_Weight' :B_Weight,
'B_Height' :B_Height,
'B_Age' :B_Age
},
result=prediction
)
@app.route('/logisticregression')
def logreg():
return(flask.render_template('00_model_logisticsregression.html'))
@app.route('/gridsearch')
def grisea():
return(flask.render_template('00_model_gridsearch.html'))
@app.route('/deeplearning')
def deelea():
return(flask.render_template('01_training_model.html'))
@app.route('/viz1')
def viz1():
return(flask.render_template('00_data_visualization_a.html'))
@app.route('/viz2')
def viz2():
return(flask.render_template('00_data_visualization_b.html'))
@app.route('/viz3')
def viz3():
return(flask.render_template('00_data_visualization_c.html'))
@app.route('/trained')
def traine():
return(flask.render_template('02_notraining_model.html'))
@app.route('/debugging')
def debugg():
return(flask.render_template('03_debugger.html'))
if __name__ == '__main__':
app.run(debug=True)
我在 heroku 中看到的错误是：
2020-05-31T19:43:05.000000+00:00 app[api]: Build succeeded
2020-05-31T19:48:29.426002+00:00 app[web.1]: 10.81.161.51 - - [31/May/2020:19:48:29 +0000] "GET / HTTP/1.1" 200 4710 "https://dashboard.heroku.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.432446+00:00 heroku[router]: at=info method=GET path="/" host=bcs-final.herokuapp.com request_id=1df55828-886a-45d1-9605-aafd3414f9e3 fwd="71.143.151.112" dyno=web.1 connect=1ms service=56ms status=200 bytes=4872 protocol=https
2020-05-31T19:48:29.623294+00:00 app[web.1]: 10.81.161.51 - - [31/May/2020:19:48:29 +0000] "GET /reset.css HTTP/1.1" 404 232 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.623792+00:00 heroku[router]: at=info method=GET path="/reset.css" host=bcs-final.herokuapp.com request_id=40816718-e41b-44e9-a3d5-da5ff2030adb fwd="71.143.151.112" dyno=web.1 connect=1ms service=7ms status=404 bytes=400 protocol=https
2020-05-31T19:48:29.625560+00:00 app[web.1]: 10.7.244.15 - - [31/May/2020:19:48:29 +0000] "GET /style.css HTTP/1.1" 404 232 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:29.627392+00:00 heroku[router]: at=info method=GET path="/style.css" host=bcs-final.herokuapp.com request_id=bd01a28c-7d1e-4dc6-9a85-405b1bdbb8ea fwd="71.143.151.112" dyno=web.1 connect=1ms service=8ms status=404 bytes=400 protocol=https
2020-05-31T19:48:37.727173+00:00 app[web.1]: [2020-05-31 19:48:37,723] ERROR in app: Exception on / [POST]
2020-05-31T19:48:37.727185+00:00 app[web.1]: Traceback (most recent call last):
2020-05-31T19:48:37.727186+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 2447, in wsgi_app
2020-05-31T19:48:37.727186+00:00 app[web.1]: response = self.full_dispatch_request()
2020-05-31T19:48:37.727187+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1952, in full_dispatch_request
2020-05-31T19:48:37.727215+00:00 app[web.1]: rv = self.handle_user_exception(e)
2020-05-31T19:48:37.727216+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1821, in handle_user_exception
2020-05-31T19:48:37.727217+00:00 app[web.1]: reraise(exc_type, exc_value, tb)
2020-05-31T19:48:37.727218+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/_compat.py", line 39, in reraise
2020-05-31T19:48:37.727218+00:00 app[web.1]: raise value
2020-05-31T19:48:37.727219+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1950, in full_dispatch_request
2020-05-31T19:48:37.727219+00:00 app[web.1]: rv = self.dispatch_request()
2020-05-31T19:48:37.727220+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/flask/app.py", line 1936, in dispatch_request
2020-05-31T19:48:37.727220+00:00 app[web.1]: return self.view_functions[rule.endpoint](**req.view_args)
2020-05-31T19:48:37.727221+00:00 app[web.1]: File "/app/app.py", line 36, in main
2020-05-31T19:48:37.727221+00:00 app[web.1]: input_scaled = scaler.transform(input_variables)
2020-05-31T19:48:37.727222+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/preprocessing/_data.py", line 794, in transform
2020-05-31T19:48:37.727222+00:00 app[web.1]: force_all_finite='allow-nan')
2020-05-31T19:48:37.727222+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/base.py", line 436, in _validate_data
2020-05-31T19:48:37.727223+00:00 app[web.1]: self._check_n_features(X, reset=reset)
2020-05-31T19:48:37.727223+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.6/site-packages/sklearn/base.py", line 373, in _check_n_features
2020-05-31T19:48:37.727223+00:00 app[web.1]: "The reset parameter is False but there is no "
2020-05-31T19:48:37.727314+00:00 app[web.1]: RuntimeError: The reset parameter is False but there is no n_features_in_ attribute. Is this estimator fitted?
2020-05-31T19:48:37.728944+00:00 app[web.1]: 10.7.244.15 - - [31/May/2020:19:48:37 +0000] "POST / HTTP/1.1" 500 290 "https://bcs-final.herokuapp.com/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36"
2020-05-31T19:48:37.730539+00:00 heroku[router]: at=info method=POST path="/" host=bcs-final.herokuapp.com request_id=2ba8ae51-f913-48d0-9b45-78aef407c15e fwd="71.143.151.112" dyno=web.1 connect=1ms service=27ms status=500 bytes=470 protocol=https
查看代码，我无法确定运行在本地和在 heroku 上使用您的应用程序之间有什么区别。唯一合理的猜测是您的本地环境与 heroku 上的 运行 版本不同。
无论如何，如果我们关注 heroku 的版本告诉我们的内容，您的错误来自这一行：
# ...
input_scaled = scaler.transform(input_variables) # <--- here
# ...
错误基本上是说：估计器 - 实际上是你的缩放器（参见源代码 here） - 不适合。
消息本身被打印出来 here 并且说“n_features_in_ 属性缺失”的部分本质上意味着您的 input_variables 稀疏矩阵（在 scikit-learn 源代码中称为 X) 缺少属性 n_features_in_。该属性设置为拟合过程的一部分。
最重要的是，在 transform 之前调用 fit，这可能会有所帮助。所以像：
input_scaled = scaler.fit(input_variables).transform(input_variables)