fullstop-punctuation-multilang-large by Oliver Guhr

Runs On

What it takes to serve fullstop-punctuation-multilang-large (559M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

Precision	Weights	Memory needed	Cheapest setup	Per hour	Also fits
16-bit	1.1 GB	1.3 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	0.6 GB	0.7 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	0.3 GB	0.3 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

Model Card

By Oliver Guhr, published under mit, revision 345e80adc07e.

This model predicts the punctuation of English, Italian, French and German texts. We developed it to restore the punctuation of transcribed spoken language.

This multilanguage model was trained on the Europarl Dataset provided by the SEPP-NLG Shared Task. Please note that this dataset consists of political speeches. Therefore the model might perform differently on texts from other domains.

The model restores the following punctuation markers: "." "," "?" "-" ":"

Sample Code

We provide a simple python package that allows you to process text of any length.

Install

To get started install the package from pypi:

pip install deepmultilingualpunctuation

Restore Punctuation

from deepmultilingualpunctuation import PunctuationModel

model = PunctuationModel()
text = "My name is Clara and I live in Berkeley California Ist das eine Frage Frau Müller"
result = model.restore_punctuation(text)
print(result)

output

My name is Clara and I live in Berkeley, California. Ist das eine Frage, Frau Müller?

Predict Labels

Read the full model card (469 words)

Configuration

Architecture: XLMRobertaForTokenClassification
Context length (tokens): 514
Layers: 24
Hidden size: 1,024
Feed-forward size: 4,096
Attention heads: 16
Vocabulary size: 250,002
Model type: xlm-roberta

Identity and Version

Repository: oliverguhr/fullstop-punctuation-multilang-large
Publisher: Oliver Guhr
Task: Token classification
Modality: Text
Library: transformers
Parameters: 559M parameters
Languages: en, de, fr, it
Revision: 345e80adc07e761d3a35feafd20f2f44a151f453
First published: 2022-03-02
Last updated: 2023-11-16

Files and Weights

18 files, 9.0 GB in total. The weights are 5 files totalling 6.7 GB in bin, h5, onnx, safetensors.

Weights5 files · 6.7 GB

Configuration4 files · 2.4 KB

Tokenizer4 files · 34.2 MB

Documentation1 file · 6.0 KB

Other3 files · 2.2 GB

Repository1 file · 1.4 KB

Every file

File	Type	Size	SHA-256
model.safetensors	Weights	2.2 GB	270f27d7398a
onnx/model.onnx	Weights	617.8 KB	8414d4e6e5b8
pytorch_model.bin	Weights	2.2 GB	b5da1fee6cc2
tf_model.h5	Weights	2.2 GB	7ab8d99474ea
training_args.bin	Weights	2.4 KB	819a2b481e39
config.json	Configuration	892 B	—
onnx/config.json	Configuration	953 B	—
onnx/special_tokens_map.json	Configuration	280 B	—
special_tokens_map.json	Configuration	239 B	—
README.md	Documentation	6.0 KB	—
onnx/model.onnx_data	Other	2.2 GB	dda38a70eda9
onnx/sentencepiece.bpe.model	Other	5.1 MB	cfc8146abe2a
sentencepiece.bpe.model	Other	5.1 MB	cfc8146abe2a
.gitattributes	Repository	1.4 KB	—
onnx/tokenizer.json	Tokenizer	17.1 MB	de3788bb2f34
onnx/tokenizer_config.json	Tokenizer	496 B	—
tokenizer.json	Tokenizer	17.1 MB	de3788bb2f34
tokenizer_config.json	Tokenizer	406 B	—

License and Download

License: mit
Access: Open weights, no gate
Download size: 6.7 GB

Download from Oliver Guhr

Released by Oliver Guhr through its official repository on Hugging Face. Read the license.

Built From

Trained on (disclosed) wmt/europarl

Memory Requirements

Precision	Weights in memory
As published	6.7 GB
16-bit	1.1 GB
8-bit	0.6 GB
4-bit	0.3 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About fullstop-punctuation-multilang-large

How much GPU memory does fullstop-punctuation-multilang-large need?

About 1.3 GB at 16-bit and 0.3 GB at 4-bit: the weights (559M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run fullstop-punctuation-multilang-large on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use fullstop-punctuation-multilang-large commercially?

Yes. fullstop-punctuation-multilang-large is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.

What is fullstop-punctuation-multilang-large's context length?

514 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Token classification

OpenMed-NER-GenomicDetect-BigMed-560M

OpenMed

Specialized model for Gene Entity Recognition - Gene-related entities This model is a state-of-the-art fine-tuned transformer engineered to deliver enterprise-grade accuracy for gene entity recognition - gene-related entities. This specialized model excels at identifying and extracting biomedical entities from clinical texts, research papers, and healthcare documents, enabling applications such as drug interaction detection, medication extraction from patient records, adverse event monitoring, literature mining for drug discovery, and biomedical knowledge graph construction with production-ready reliability for clinical and research applications. This model can identify and classify the…

Open weights apache-2.0 559M parameters 514 tokens transformers

View model

Model · Token classification

OpenMed-NER-ChemicalDetect-BigMed-560M

OpenMed

Specialized model for Chemical Entity Recognition - Identifies chemical compounds and substances in biomedical literature This model is a state-of-the-art fine-tuned transformer engineered to deliver enterprise-grade accuracy for chemical entity recognition - identifies chemical compounds and substances in biomedical literature. This specialized model excels at identifying and extracting biomedical entities from clinical texts, research papers, and healthcare documents, enabling applications such as drug interaction detection, medication extraction from patient records, adverse event monitoring, literature mining for drug discovery, and biomedical knowledge graph construction with…

Open weights apache-2.0 559M parameters 514 tokens transformers

View model

Model · Token classification

OpenMed-NER-BloodCancerDetect-ElectraMed-560M

OpenMed

Specialized model for Clinical Entity Recognition - Clinical entities related to Chronic Lymphocytic Leukemia This model is a state-of-the-art fine-tuned transformer engineered to deliver enterprise-grade accuracy for clinical entity recognition - clinical entities related to chronic lymphocytic leukemia. This specialized model excels at identifying and extracting biomedical entities from clinical texts, research papers, and healthcare documents, enabling applications such as drug interaction detection, medication extraction from patient records, adverse event monitoring, literature mining for drug discovery, and biomedical knowledge graph construction with production-ready reliability for…

Open weights apache-2.0 559M parameters 514 tokens transformers

View model

Model · Token classification

llmlingua-2-xlm-roberta-large-meetingbank

Microsoft

This model was introduced in the paper LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression (Pan et al, 2024). It is a XLM-RoBERTa (large-sized model) finetuned to perform token classification for task agnostic prompt compression. The probability $p{preserve}$ of each token $xi$ is used as the metric for compression. This model is trained on the extractive text compression dataset constructed with the methodology proposed in the LLMLingua-2, using training examples from MeetingBank (Hu et al, 2023) as the seed data. You can evaluate the model on downstream tasks such as question answering (QA) and summarization over compressed meeting transcripts using…

Open weights mit 559M parameters 514 tokens transformers

View model

Model · Token classification

OpenMed-NER-ChemicalDetect-MultiMed-568M

OpenMed

Specialized model for Chemical Entity Recognition - Identifies chemical compounds and substances in biomedical literature This model is a state-of-the-art fine-tuned transformer engineered to deliver enterprise-grade accuracy for chemical entity recognition - identifies chemical compounds and substances in biomedical literature. This specialized model excels at identifying and extracting biomedical entities from clinical texts, research papers, and healthcare documents, enabling applications such as drug interaction detection, medication extraction from patient records, adverse event monitoring, literature mining for drug discovery, and biomedical knowledge graph construction with…

Open weights apache-2.0 567M parameters 8,194 tokens transformers

View model

Model · Token classification

OpenMed-NER-OncologyDetect-MultiMed-568M

OpenMed

Specialized model for Cancer Genetics - Cancer-related genetic entities This model is a state-of-the-art fine-tuned transformer engineered to deliver enterprise-grade accuracy for cancer genetics - cancer-related genetic entities. This specialized model excels at identifying and extracting biomedical entities from clinical texts, research papers, and healthcare documents, enabling applications such as drug interaction detection, medication extraction from patient records, adverse event monitoring, literature mining for drug discovery, and biomedical knowledge graph construction with production-ready reliability for clinical and research applications. This model can identify and classify the…

Open weights apache-2.0 567M parameters 8,194 tokens transformers

View model