Muhammad4hmed

from scipy.optimize import minimize

#alloff contains 5 fold prediction on all dataset (avoid data leak)

def min_func(K):
    ypredtrain = 0
    for a in range(len(alloof)):

        ypredtrain += K[a]*alloof[a]

Muhammad4hmed

https://www.kaggle.com/code/cdeotte/forward-selection-oof-ensemble-0-942-private/notebook

Muhammad4hmed

Finally we have more than 1000 features in all.

Ok lets get a little bit more technical:
Overview of solution
Preprocessing

1.remove punctuations

2.using porter stemmer

Muhammad4hmed

import numpy as np 
import pandas as pd 

import os
from transformers import *

from sklearn.model_selection import KFold, StratifiedKFold
from sklearn.metrics import mean_squared_error, accuracy_score, roc_auc_score

import random

Muhammad4hmed

class Squeeze(nn.Module):
    def __init__(self, dims=-1):
        super().__init__()
        self.dims = dims
        
    def forward(self, x):
        return x.squeeze(self.dims)
        
class AttentionHead(nn.Module):
    def __init__(self, in_features, hidden_dim, num_targets):

Muhammad4hmed

def groupby(train, test, col):
    res = train.groupby(col)['Price'].agg(['min','mean','max'])
    
    train[f'{col}_Min'] = train[col].map(res['min'])
    train[f'{col}_Mean'] = train[col].map(res['mean'])
    train[f'{col}_Max'] = train[col].map(res['max'])
    
    test[f'{col}_Min'] = test[col].map(res['min'])
    test[f'{col}_Mean'] = test[col].map(res['mean'])
    test[f'{col}_Max'] = test[col].map(res['max'])

Muhammad4hmed

def extract_w_h(image_id_check):
    # Check weight height order
    w, h = test_df.loc[test_df.id==image_id,['dim1', 'dim0']].values[0]
    return w, h

def extract_data(data_list, w, h):
    
    labels = []
    scores_list=[]
    boxes_list = []

Muhammad4hmed

https://www.kaggle.com/sreevishnudamodaran/siim-mmdetection-cascadercnn-weight-bias/notebook

Muhammad4hmed

class LitModel(nn.Module):
    def __init__(self):
        super().__init__()

        config = AutoConfig.from_pretrained(ROBERTA_PATH)
        config.update({"output_hidden_states":True, 
                       "hidden_dropout_prob": 0.0,
                       "layer_norm_eps": 1e-7})                       
        
        self.roberta = AutoModel.from_pretrained(ROBERTA_PATH, config=config)  

Muhammad4hmed

import pandas as pd

import warnings
warnings.filterwarnings('ignore')

from transformers import (AutoModel,AutoModelForMaskedLM, 
                          AutoTokenizer, LineByLineTextDataset,
                          DataCollatorForLanguageModeling,
                          Trainer, TrainingArguments)
train_data = pd.read_csv('../input/commonlitreadabilityprize/train.csv')