Hello dev.Community! 👋 In this post, I'll guide you through the process of image classification using JavaScript with the assistance of Hugging Face AI Models.
First and foremost, let's select a random image online and download it. For example, I came across this picture during a quick Google search. Once we have our chosen image, let's open Visual Studio Code (VSCode) and initiate our Node.js project by executing the following command in the terminal:
After running this command, a package.json file will be generated.
Now, let's get down to business and install our Hugging Face model. If you're not familiar with
Hugging Face
It's a company renowned for its Natural Language Processing (NLP) platform and an extensive open-source community that offers various pre-trained models, including transformers for tasks like text generation, translation, and sentiment analysis.
To install our specific model, Xenova/transformers.js, use the following terminal command:*
Xenova/transformers.js
with this command in terminal:
To install our specific model, Xenova/transformers.js, use the following terminal command:
npm install Xenova/transformers.js --save
Upon installing the model, a dependency entry will be added to the package.json file, as depicted below:
Next, modify the package.json by adding "type": "module" because we want to leverage ES6 features.
Next step we create an index.js file in our project folder , you need to write these lines of code :
Here we imported pipeline from Xenova/transformers.js then initialize a pipe to use with an async function to tell what this pipeline is for also the task we want to accomplish is we want to make an image classification , you can specify also a specific model you want to use but here in this post we just stick with the default one cause it will quicker to download , we declared also a const called result so we can process an image so we need to provide an absolute path for the image,finally we just console.log the result.
Run Node index in terminal :
So as shown in terminal label it tell us that is a lion king and a score of possibiliy with a high percantage 99% but if we want our result to be more accurate add the option topk:6 to the result constant.
here it shows other labels with 0% percantage so defiently it is a lion
🦁!
Top comments (2)
Hello dev.Community! 😊
Excited to share my very first post with all of you. A big thanks to the community for this opportunity! In this post, I delve into the fascinating world of image classification with JavaScript and Hugging Face AI Models. Would love for you to check it out, share your thoughts, and feel free to drop any questions or feedback. Looking forward to engaging with this awesome community. Happy coding!