Bitfount-Supported Algorithms

Algorithms dictate what type of task a Data Scientist is looking to execute. Bitfount requires specification of these algorithms to ensure a Data Scientist has the appropriate authorisation to execute them.

Bitfount currently supports the following algorithms, which are described in more technical detail in the API Reference:

Algorithm	Description
Federated Training	Algorithm for training a model remotely and returning its updated parameters.
Evaluate	Algorithm for evaluating a model and returning metrics.
Train and Evaluate	Algorithm for training a model, evaluating it, and returning metrics.
Inference	Algorithm for running inference on a model and returning predictions.
SQL Query	Simple algorithm for running a SQL query on a table.
Private SQL Query	Enables the user to run a private SQL query using differential privacy by specifying epsilon and delta.
CSV Report	Saves the results of another algorithm as a CSV on the Pod.
Hugging Face Transformer Perplexity	Computes and returns the perplexity of a huggingface model on the Pod's text data.
Hugging Face Transformer Text Generation	Generates text using a huggingface model. Prompts can exist on the Pod or be sent dynamically by the Data Scientist.
Hugging Face Image Classification	Image Classification Algorithm for pre-trained HuggingFace models.
Hugging Face Image Segmentation	Image Segmentation Algorithm for pre-trained HuggingFace models.
Hugging Face Text Classification	Text Classification Algorithm for pre-trained HuggingFace models.
Hugging Face TIMM Fine Tuning	Algorithm for fine-tuning Hugging Face image models from the TIMM library.
Hugging Face TIMM Inference	Algorithm for running inference on Hugging Face image models from the TIMM library.

Don’t see an algorithm that works for you? Submit your suggestion or provide feedback to us.