T5 Small Openvino

🚀 t5-small Exported to the OpenVINO IR

This project focuses on exporting the t5-small model to the OpenVINO Intermediate Representation (IR). It enables tasks like summarization and translation, leveraging the power of OpenVINO.

🚀 Quick Start

The model can be easily used with the Transformers pipeline. Here's a simple example to get you started.

✨ Features

• Multilingual Support: Supports languages such as English, French, Romanian, and German.
• Diverse Datasets: Trained on the c4 dataset.
• Task Capabilities: Capable of tasks like summarization, translation, and is optimized with OpenVINO.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, pipeline
from optimum.intel.openvino import OVModelForSeq2SeqLM

model_id = "echarlaix/t5-small-openvino"
model = OVModelForSeq2SeqLM.from_pretrained(model_id, use_cache=False)
tokenizer = AutoTokenizer.from_pretrained(model_id)

# Create a pipeline
translation_pipe = pipeline("translation_en_to_fr", model=model, tokenizer=tokenizer)

text = "He never went out without a book under his arm, and he often came back with two."
result = translation_pipe(text)

📚 Documentation

Model description

T5 is an encoder-decoder model pre-trained on a multi-task mixture of unsupervised and supervised tasks and for which each task is converted into a text-to-text format.

For more information, please take a look at the original paper.

Paper: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

Authors: Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu

Information Table

📄 License

This project is licensed under the Apache-2.0 license.