i need add some wat to increse performnce for this code please : import pandas as pd

from keras.models import Sequential

from keras.layers import Dense

from keras.preprocessing.text import Tokenizer

from sklearn.preprocessing import LabelEncoder

from keras.utils import to$\_$categorical

from sklearn.metrics import f$1\_$score

import numpy as np

# Load training data

df$\_$train $=$ pd$.$read$\_$csv$("$Last$\_$Complete$\_$Data$\_8224.$csv$")$

# Define your Keras model

sentiment$\_$model $=$ Sequential$()$

sentiment$\_$model.add$($Dense$(512,$ input$\_$shape$=(6000,),$ activation$=$'relu'$))$

sentiment$\_$model.add$($Dense$(256,$ activation$=$'relu'$))$

sentiment$\_$model.add$($Dense$(128,$ activation$=$'relu'$))$

sentiment$\_$model.add$($Dense$(64,$ activation$=$'relu'$))$

sentiment$\_$model.add$($Dense$(32,$ activation$=$'relu'$))$

sentiment$\_$model.add$($Dense$(2,$ activation$=$'sigmoid'$))$

# Compile the model

sentiment$\_$model.compile$($loss$=$'categorical$\_$crossentropy', optimizer$=$'Adam', metrics$=[\text{'}$accuracy$\text{'}])$

# Tokenize text data

vocab$\_$size $=6000$

tokenizer $=$ Tokenizer$($num$\_$words$=$vocab$\_$size$)$

tokenizer.fit$\_$on$\_$texts$($df$\_$train$[\text{'}$Pre$\_$Text'$])$

train$\_$reviews$\_$tokenized $=$ pd$.$DataFrame$($tokenizer$.$texts$\_$to$\_$matrix$($df$\_$train$[\text{'}$Pre$\_$Text'$]))$

# Encode the label variable

label$\_$encoder $=$ LabelEncoder$()$

integer$\_$sentiment $=$ label$\_$encoder.fit$\_$transform$($df$\_$train$[\text{'}$sentiment$\text{'}])$

encoded$\_$y$\_$train $=$ to$\_$categorical$($integer$\_$sentiment$)$

# Train the model

for epoch in range$(10)$:

history $=$ sentiment$\_$model.fit$($train$\_$reviews$\_$tokenized, encoded$\_$y$\_$train, epochs$=1,$ validation$\_$split$=0\mathrm{.}2,$ verbose$=1)$

# Calculate training accuracy

train$\_$accuracy $=$ history.history$[\text{'}$accuracy$\text{'}][0]$

# Predict classes for validation set

val$\_$predictions $=$ sentiment$\_$model.predict$($train$\_$reviews$\_$tokenized$)$

val$\_$predicted$\_$classes $=$ np$.$argmax$($val$\_$predictions, axis$=1)$

# Calculate F$1-$score on validation set

val$\_$f$1=$ f$1\_$score$($integer$\_$sentiment$,$ val$\_$predicted$\_$classes, average$=$'weighted'$)$

print$("$Epoch $\{\}$: Training Accuracy $=\{$:$\mathrm{.}4$f$\},$ Validation F$1-$score $=\{$:$\mathrm{.}4$f$\}".$format$($epoch$+1,$ train$\_$accuracy, val$\_$f$1))$