Model Training Basics
When training models using AI Studio, compute runs locally on the user's laptop. Once users have access to AI Studio, models can be trained through the notebooks section. Users can write code, collaborate with team members, run the code and track model training experiments through MLFlow and TensorBoard dashboards.
AI Studio Effect on Model Accuracy
Artificial Intelligence and Machine Learning solutions can only ever be as accurate as the data the solution is trained on. AI Studio will not independently alter or fact check the data provided. However, many of the tools included in AI Studio do promote good data hygiene practices which can have a significant effect on the reliability of the model created.
NLP/LLM libraries included with Large Image
Large images include Tensorflow and Pytorch, common packages for running Deep Learning experiments as well as Scikit-learn for basic NLP. AI Studio does not have an NLP/LLM base image at this time; however, the current version allows users to create, activate, and install custom Conda environments for a project that persist, which allow users to define once and reuse on their system in the future.