superduper-transformers 0.4.0

Transformers is a popular AI framework, and we have incorporated native support for Transformers to provide essential Large Language Model (LLM) capabilities.

superduper_transformers

Transformers is a popular AI framework, and we have incorporated native support for Transformers to provide essential Large Language Model (LLM) capabilities.

Superduper allows users to work with arbitrary transformers pipelines, with custom input/ output data-types.

Installation

pip install superduper_transformers

API

• Code
• API-docs

Class	Description
superduper_transformers.model.TextClassificationPipeline	A wrapper for transformers.Pipeline.
superduper_transformers.model.LLM	LLM model based on transformers library.

Examples

TextClassificationPipeline

from superduper_transformers.model import TextClassificationPipeline

model = TextClassificationPipeline(
    identifier="my-sentiment-analysis",
    model_name="distilbert-base-uncased",
    model_kwargs={"num_labels": 2},
    device="cpu",
)
model.predict("Hello, world!")

LLM

from superduper_transformers import LLM
model = LLM(identifier="llm", model_name_or_path="facebook/opt-125m")
model.predict("Hello, world!")

Training Example

Read more about this here