How to convert .dat to a torch.tensor in python

I have a DAT file with the following structure:

;Text1
;Text2
;Text3

;    N    Vec1       Vec2    

     1  -0.09999   0.09999
     2  -0.09999  -0.09999 
     3  -0.09999  -0.09999 

and I want to store the values of Vec1 and Vec2 into two torch tensors.

I tried to use pandas:

import torch
import pandas as pd
df = pd.read_csv('file.DAT',skiprows=4, encoding='latin-1')
vec1 = torch.tensor(df.iloc[:,1].values,  dtype=torch.float32)
vec2 = torch.tensor(df.iloc[:,2].values,  dtype=torch.float32)

But the problem is I cannot reach the values inside the resulting frame "df". Can someone help? I just want to store Vec1 and Vec2 into two torch tensors.

>Solution :

You can use read_fwf :

df = pd.read_fwf("file.DAT", skiprows=3) #`skip_blank_lines=True` by default

vec1 = torch.tensor(df["Vec1"], dtype=torch.float32)
vec2 = torch.tensor(df["Vec2"], dtype=torch.float32)

Output :

print(vec1) #tensor([-0.1000, -0.1000, -0.1000])
print(vec2) #tensor([ 0.1000, -0.1000, -0.1000])