models t5 large

t5-large

Overview

The T5-Large is a text-to-text transfer transformer (T5) model with 770 million parameters. It has been developed by Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu. The T5 model is a language model that is pre-trained on a multi-task mixture of unsupervised and supervised tasks. The model is versatile and can be used for many NLP tasks such as translation, summarization, question answering, and classification. The license of T5-Large is Apache 2.0. The model is made available on GitHub and is well-documented on Hugging Face. A code sample is provided to get started with this model. The input and output are always text strings. The T5 framework was introduced to bring together transfer learning techniques for NLP and convert all language problems into the text to text format. The T5 model was pre-trained on the Colossal Clean Crawled Corpus (C4). The full details of the training procedure can be found in the research paper. The evaluation of T5-Large is not provided. The information regarding bias, risks, limitations, and recommendations is not available. The authors of the Model Card for T5-Large are not specified.

The above summary was generated using ChatGPT. Review the original-model-card to understand the data used to train the model, evaluation metrics, license, intended uses, limitations and bias before using the model.

Inference samples

Inference type	Python sample (Notebook)	CLI with YAML
Real time	translation-online-endpoint.ipynb	translation-online-endpoint.sh
Batch	translation-batch-endpoint.ipynb	coming soon

Finetuning samples

Task	Use case	Dataset	Python sample (Notebook)	CLI with YAML
Summarization	News Summary	CNN DailyMail	news-summary.ipynb	news-summary.sh
Translation	Translate English to Romanian	WMT16	translate-english-to-romanian.ipynb	translate-english-to-romanian.sh

Model Evaluation

Task	Use case	Dataset	Python sample (Notebook)	CLI with YAML
Translation	Translation	wmt16/ro-en	evaluate-model-translation.ipynb	evaluate-model-translation.yml

Sample inputs and outputs (for real-time inference)

Sample input

{
    "inputs": {
        "input_string": ["My name is John and I live in Seattle", "Berlin is the capital of Germany."]
    },
    "parameters": {
        "task_type": "translation_en_to_fr"
    }
}

Sample output

[
    {
        "0": "Mon nom est John et je vivais à Seattle."
    },
    {
        "0": "Berlin est la capitale de l'Allemagne."
    }
]

Version: 11

Tags

Preview computes_allow_list : ['Standard_NC6s_v3', 'Standard_NC12s_v3', 'Standard_NC24s_v3', 'Standard_NC24rs_v3', 'Standard_NC6s_v2', 'Standard_NC12s_v2', 'Standard_NC24s_v2', 'Standard_NC24rs_v2', 'Standard_NC4as_T4_v3', 'Standard_NC8as_T4_v3', 'Standard_NC16as_T4_v3', 'Standard_NC64as_T4_v3', 'Standard_ND6s', 'Standard_ND12s', 'Standard_ND24s', 'Standard_ND24rs', 'Standard_ND40rs_v2', 'Standard_ND96asr_v4'] license : apache-2.0 model_specific_defaults : ordereddict([('apply_deepspeed', 'true'), ('apply_lora', 'true'), ('apply_ort', 'true')]) task : text-translation

View in Studio: https://ml.azure.com/registries/azureml/models/t5-large/version/11

License: apache-2.0

Properties

SHA: 6ed159b7a9408621a1ed0b9ca39aacc42c694d73

datasets: c4

evaluation-min-sku-spec: 8|0|28|56

evaluation-recommended-sku: Standard_DS4_v2

finetune-min-sku-spec: 4|1|28|176

finetune-recommended-sku: Standard_NC24rs_v3

finetuning-tasks: summarization, translation

inference-min-sku-spec: 8|0|28|56

inference-recommended-sku: Standard_DS4_v2, Standard_D8a_v4, Standard_D8as_v4, Standard_DS5_v2, Standard_D16a_v4, Standard_D16as_v4, Standard_D32a_v4, Standard_D32as_v4, Standard_D48a_v4, Standard_D48as_v4, Standard_D64a_v4, Standard_D64as_v4, Standard_D96a_v4, Standard_D96as_v4, Standard_FX12mds, Standard_F16s_v2, Standard_F32s_v2, Standard_F48s_v2, Standard_F64s_v2, Standard_F72s_v2, Standard_FX24mds, Standard_FX36mds, Standard_FX48mds, Standard_E8s_v3, Standard_E16s_v3, Standard_E32s_v3, Standard_E48s_v3, Standard_E64s_v3, Standard_NC4as_T4_v3, Standard_NC6s_v3, Standard_NC8as_T4_v3, Standard_NC12s_v3, Standard_NC16as_T4_v3, Standard_NC24s_v3, Standard_NC64as_T4_v3, Standard_NC24ads_A100_v4, Standard_NC48ads_A100_v4, Standard_NC96ads_A100_v4, Standard_ND96asr_v4, Standard_ND96amsr_A100_v4, Standard_ND40rs_v2

languages: en, fr, ro, de