Hello my C1W4_assignment has it’s error, anyone can help me?
1 Like
Dear @Shandi_Adhitya,
Welcome to the community.
Can you please send me a private message with your C1_W4_Assignment.js as attachment.
I’ll look into the issues, let you know the solution.
Hi @Girijesh @Shandi_Adhitya , im facing the same issue while trying to build my tensorflow model, it’ll be greate if you could help me out with it!
1 Like
Please send me your notebook file to me. And also let me know me know the problems that are you facing.
I am actually trying to implement a model in a react native application, and im using these tensorflow libraries : @tensorflow/tfjs, @tensorflow/tfjs-react-native.I’m not able to attach my github repo link here for some reason.
This is the error that im encountering in my sentimentAnalysis.js file
ERROR Error during prediction: [TypeError: Cannot read property ‘isTypedArray’ of undefined]
and this is specifically occuring in the const input line of the below code snippet
import React, { useState, useEffect } from 'react';
import { View, Text, TextInput, Button, ActivityIndicator } from 'react-native';
import * as tf from '@tensorflow/tfjs';
import '@tensorflow/tfjs-react-native';
const SentimentAnalysis = () => {
const [model, setModel] = useState(null);
const [indexFrom, setIndexFrom] = useState(null);
const [maxLen, setMaxLen] = useState(null);
const [wordIndex, setWordIndex] = useState(null);
const [vocabularySize, setVocabularySize] = useState(null);
const [searchText, setSearchText] = useState('');
const [loading, setLoading] = useState(true);
useEffect(() => {
const loadModelAndMetadata = async () => {
try {
console.log('Loading model');
// Wait for TensorFlow.js to be ready
await tf.ready();
// Set the fetch function to the default fetch provided by React Native
tf.io.registerLoadRouter((url) => {
if (url.startsWith('https://')) {
return fetch(url, { method: 'GET' }).then((response) => response.arrayBuffer());
}
return null;
});
const modelUrl = 'https://storage.googleapis.com/tfjs-models/tfjs/sentiment_cnn_v1/model.json';
const loadedModel = await fetch(modelUrl)
.then((response) => response.json())
.then((model) => tf.loadLayersModel(tf.io.fromMemory(model)))
.catch((error) => {
console.error('Error loading model:', error);
throw error;
});
setModel(loadedModel);
const sentimentMetadata = await getMeta('https://storage.googleapis.com/tfjs-models/tfjs/sentiment_cnn_v1/metadata.json');
setIndexFrom(sentimentMetadata['index_from']);
setMaxLen(sentimentMetadata['max_len']);
setWordIndex(sentimentMetadata['word_index']);
setVocabularySize(sentimentMetadata['vocabulary_size']);
setLoading(false); // Set loading to false once everything is loaded
console.log('Loaded model');
} catch (error) {
console.error('Error loading model or metadata:', error);
setLoading(false); // Set loading to false in case of an error
}
};
loadModelAndMetadata();
}, []);
const padSequences = (sequence, maxLen, padding = 'pre', truncating = 'pre', value = 0) => {
let paddedSequence = sequence.slice(0, maxLen);
if (padding === 'pre') {
paddedSequence = Array.from({ length: maxLen - sequence.length }, () => value).concat(paddedSequence);
} else {
paddedSequence = paddedSequence.concat(Array.from({ length: maxLen - sequence.length }, () => value));
}
return paddedSequence;
};
const predictions = async () => {
if (loading) {
console.error('Model or metadata is still loading.');
return;
}
if (!model || !indexFrom || !maxLen || !wordIndex || !vocabularySize) {
console.error('Model or metadata not loaded yet.');
return;
}
const inputText = searchText.trim().toLowerCase().replace(/(\.|\,|\!)/g, '').split(' ');
console.log('Input Text:', inputText);
const sequence = inputText.map((word) => {
console.log('word:', word);
let wordIndexValue = wordIndex[word] + indexFrom;
if (wordIndexValue > vocabularySize) {
wordIndexValue = 2;
}
console.log('wordIndexValue:', wordIndexValue);
return wordIndexValue;
});
const paddedSequence = padSequences(sequence, maxLen);
console.log('Padded Sequence:', paddedSequence, typeof(paddedSequence));
console.log('Padded Sequence Length:', paddedSequence.length);
// Convert the paddedSequence to Int32Array
const inputArray = Int32Array.from(paddedSequence);
// Create a tensor from the Int32Array
// Wrap the prediction code in a try-catch block to handle errors
try {
const input = tf.tensor([Array.from(inputArray)], [1, maxLen], 'int32'); // Fix here
console.log('Input Tensor:', input);
console.log('Input Shape:', input.shape); // Log the shape for debugging
const beginMs = performance.now();
const predictOut = model.predict(input); // Fix here
const score = predictOut.dataSync()[0];
predictOut.dispose();
const endMs = performance.now();
console.log('Predicted Score:', score);
console.log('Prediction successful');
} catch (error) {
console.error('Error during prediction:', error);
}
};
const getMeta = async (path) => {
const response = await fetch(path);
const data = await response.json();
return data;
};
return (
<View>
<Text>Input Text:</Text>
<TextInput
style={{ height: 40, borderColor: 'gray', borderWidth: 1, marginBottom: 10 }}
onChangeText={(text) => setSearchText(text)}
value={searchText}
/>
<Button title="Predict Sentiment" onPress={predictions} />
{loading && <ActivityIndicator size="large" color="#0000" />}
</View>
);
};
export default SentimentAnalysis;
1 Like
Dear @Anirudh13,
Please send me all the required files via personal message.
Click on my profile image and click on message.
I’ll look into the issues and let you know the solutions.
Did you find answer to you question?
1 Like
Dear @TAYAB_FAROOQ,
Let me know the issues that are you facing.
1 Like
Dear @jahedulsumon32,
Can you please send me your notebook as personal message to me.
I’ll look into the issues and let you know the solution.