You are here:
Monitor Model Activity
View your model inference use and errors in a dashboard. Use important indicators to help you track, adjust, and improve model performance.
To monitor the performance of a model that you configured, go to AI Models (formerly Einstein Studio) Model Home. Select the model and refer to the Model Activity tab.
Inferences
A foundation model is pretrained on large datasets, but it also learns from prompts during inference when generating human-like text or performing specific language-related tasks. It’s important to adapt your model and fine-tune hyperparameters for domain-specific requirements. When doing so, factor in model efficiency, deployment infrastructure, and ethical implications.
The model activity provides a snapshot of total inferences from all configured models that you used in a specific period.
Errors
Errors can occur during the operation and deployment of the model. The errors can impact the performance, reliability, and security of an LLM in real-world applications. Common issues include inaccurate or biased responses, latency, and scalability issues.
