Supported code editor & framework
What Can I Do With Code Editor?
Jupyter Lab
A code editor on AI Platforms, like Jupyter Lab, offers a robust environment for data exploration, analysis, and model development in various programming languages like Python, R, and Julia. Here's what you can do with a code editor like Jupyter Lab:
Interactive Computing: Perform interactive computing by running code cells individually or collectively. Jupyter Lab allows for executing code in sections, making it easy to test and debug small portions of your code.
Data Visualization: Create visualizations directly within the code editor using libraries like Matplotlib, Seaborn, and Plotly. Visualizing data is crucial for data analysis, model evaluation, and gaining insights from your datasets.
Data Cleaning and Preprocessing: Use Python or other programming languages supported by the environment to clean and preprocess datasets. You can perform tasks like handling missing values, encoding categorical data, and scaling numerical features.
Model Development and Training: Develop machine learning models using popular libraries like TensorFlow and PyTorch. You can write code to define, train, evaluate, and optimize models for various tasks like classification, regression, clustering, etc.
Experimentation and Prototyping: Experiment with different algorithms, hyperparameters, and model architectures to prototype and find the best approach for your machine learning or data analysis tasks.
Documentation and Collaboration: Create comprehensive documentation using Markdown cells to explain code logic, instructions, or findings within the same environment. Collaborate with team members by sharing Jupyter notebooks.
Integration with External Services: Integrate Jupyter Lab with external services like cloud storage (e.g., S3 buckets), databases, APIs, or other AI services to access and analyze data or perform computations using external resources.
Resource Monitoring: Monitor resource utilization within the Jupyter Lab environment, including CPU, memory, and storage, to ensure efficient use of computational resources.
Extension and Customization: Customize your Jupyter Lab environment by installing extensions, and widgets, or integrating additional functionalities and libraries based on your specific requirements.
Overall, Jupyter Lab and similar code editors provide a versatile and flexible environment for performing various data-related tasks, allowing data scientists, researchers, and developers to efficiently analyze data, build machine learning models, and prototype solutions.
Please refer https://jupyterlab.readthedocs.io/en/latest/ to know more about Jupyter Lab.
Which framework I should use?
PyTorch and TensorFlow are popular deep learning frameworks used with Jupyter on the AI Platform. Both frameworks have their strengths and weaknesses, and the choice between PyTorch and TensorFlow often depends on specific project requirements, familiarity, and personal preferences of users. Jupyter Notebooks provide an interactive and intuitive interface for working with both PyTorch and TensorFlow, allowing users to write, test, and visualize their machine-learning models in a flexible environment. Here are some differences between the two:
Criteria/Framework | Pytorch | TensorFlow |
|---|---|---|
Programming Style | Known for its dynamic computational graph, PyTorch offers a more Pythonic and imperative programming style, making it easier to write and debug code due to its similarity to Python. | Traditionally uses a static computational graph paradigm with TensorFlow 1.x (but offers eager execution in TensorFlow 2.x), which requires defining the entire computation graph before running it. This approach can offer better performance optimization in some cases but may be less intuitive for beginners. |
Ease of Use | Often praised for its simplicity and ease of use, making it more beginner-friendly and allowing for quicker prototyping and experimentation. | Historically had a steeper learning curve due to its static graph nature. However, TensorFlow 2.x's adoption of eager execution has made it more user-friendly and similar in style to PyTorch. |
Community and Ecosystem | Has a larger user base and well-established community support. TensorFlow offers extensive documentation, tutorials, and pre-trained models through TensorFlow Hub. | Although growing rapidly, its community may be slightly smaller in comparison. However, it's known for its active research community, especially in the academic and research domains. |
Deployment and Production | Historically more prevalent in production environments and industrial applications, with comprehensive support for deployment on various platforms like TensorFlow Serving and TensorFlow Lite for mobile. | While gaining traction in deployment solutions, it has been more commonly associated with research and prototyping. Deployment solutions might not be as mature compared to TensorFlow. |
Visualization and Debugging | TensorFlow offers TensorBoard, a visualization toolkit for TensorFlow to track and visualize metrics, model graphs, and more. | Has libraries like TensorBoardX that allow users to log data to visualize training and loss curves, though it might not be as comprehensive as TensorBoard. |
Related Articles
Supported Models
GreenNode MaaS API provides full-precision access to the DeepSeek, Llama, and Qwen model families listed below. Model Details GreenNode MaaS supports several model types: Text generation Model Name Provider Quantization Context Length View on ...Create a notebook instance with pre-built container
A pre-built notebook instance is a managed environment where you can easily run Jupyter notebooks with access to Greennode AI resources. These instances are pre-configured with the computational resources you choose and can be used for data analysis, ...Create a notebook instance with custom container
Custom images allow you to create a notebook environment that is precisely tailored to your specific needs. You can pre-install specific libraries, frameworks, or tools that are not included in the standard application-supported images. This saves ...Playground
The GreenNode Playground is a web-based interface that allows you to test and compare AI models quickly and intuitively—without writing a single line of code. Designed for both developers and AI enthusiasts, it supports a wide range of ...Connect to a notebook instance
Greennode AI Platform offers three methods to connect to notebook instances: Code Editor: Directly access and edit code within the platform's web-based code editor. TCP Port: Connect to the notebook instance using a specified TCP port. SSH: Connect ...
