abhishekkrthakur / mlframework Goto Github PK

Hi Abhishek,
I was watching your mlframework videos (thanks for them btw!) and coding along, when I saw that this repo doesn't have a license. I wanted to use this framework for some of my own projects, so I was wondering if you could add one if possible?

Hi All,
I am trying to replicate this environment in my system. I was following the steps indicated in the book. Approaching any Machine Learning Problem.
I tried running the following the command and was receiving the following error.
I have never worked with Anaconda.
It would be great if someone can help me out.

https://github.com/scikit-learn-contrib/category_encoders

don't you think it will better to encode each column with his category type for the contest ?

I am following your YouTube series....and this happens in episode 1. I rechecked script 10 times. It is ditto same as yours.

When I ran the create_folds.py script I got uneven folds, structurally unlike what you obtained. I didn't change your code so I am not sure what is the underlying problem.

The first and last folds have this one hanging.

How to use CrossValidation functionality in train or predict implementation files?

At time stamp 11:14 in the youtube video: Tips N Tricks #2: Setting up development environment for machine learning, issuing the command conda env create -f environment.yml throws the following error.

Solving environment: failed
ResolvePackageNotFound:
  - python==3.7.6=h0371630_1

• Using Windows Subsystem as Linux (WSL) - Ubuntu

File "../src/train.py", line 12, in
FOLD = int(os.environ.get("FOLD"))
TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

I am having this issue.
Can anyone provide any solution?
TIA

After executing this command "sh run.sh randomforest", I got the following error messages.

Traceback (most recent call last):
File "/opt/anaconda3/envs/kaggle/lib/python3.8/runpy.py", line 193, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/opt/anaconda3/envs/kaggle/lib/python3.8/runpy.py", line 86, in _run_code
exec(code, run_globals)
File "/Users/ekaratrattagan/Program/Course/machine_learning/kaggle/e01/src/train.py", line 45, in
train_df.loc[:, c] = lbl.transform(train_df[c].values.tolist())
File "/opt/anaconda3/envs/kaggle/lib/python3.8/site-packages/sklearn/preprocessing/_label.py", line 273, in transform
_, y = encode(y, uniques=self.classes, encode=True)
File "/opt/anaconda3/envs/kaggle/lib/python3.8/site-packages/sklearn/preprocessing/_label.py", line 117, in _encode
return _encode_numpy(values, uniques, encode,
File "/opt/anaconda3/envs/kaggle/lib/python3.8/site-packages/sklearn/preprocessing/_label.py", line 49, in _encode_numpy
raise ValueError("y contains previously unseen labels: %s"
ValueError: y contains previously unseen labels: [nan, nan, nan, nan, nan, nan, nan, nan, ....

I then fixed it by adding the following two lines,
train_df[c].replace(np.nan, 'NAN', inplace=True)
valid_df[c].replace(np.nan, 'NAN', inplace=True)
,after for c in train_df.columns: and before lbl = preprocessing.LabelEncoder()
in train.py

label_encoders = {}
for c in train_df.columns:
train_df[c].replace(np.nan, 'NAN', inplace=True)
valid_df[c].replace(np.nan, 'NAN', inplace=True)
lbl = preprocessing.LabelEncoder()

After that, it worked perfectly.