SAVRN
Search Contact SAVRN

Open-weight model · Summarization

FRED-T5-Summarizer

by Natural Language Processing in Russian RussianNLP/FRED-T5-Summarizer

The model was trained by SberDevices. The model is trained on a mix of open summarisation data RussianNLP/Mixed-Summarization-Dataset for the Russian language and use prefix tokenen '\ '

Parameters1.7B
Context
Weights7.0 GB
Licensemit
AccessOpen weights
Monthly Downloads2.1k

Runs On

What it takes to serve FRED-T5-Summarizer (1.7B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 3.5 GB 4.2 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 1.7 GB 2.1 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
4-bit 0.9 GB 1.0 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 Natural Language Processing in Russian, published under mit, revision cb1c9278a394.

The model was trained by SberDevices. The model is trained on a mix of open summarisation data RussianNLP/Mixed-Summarization-Dataset for the Russian language and use prefix tokenen '\ '

Read Natural Language Processing in Russian's full model card

FRED-T5 1.7B Summarizer

The model was trained by SberDevices.

The model is trained on a mix of open summarisation data RussianNLP/Mixed-Summarization-Dataset for the Russian language and use prefix tokenen '\<LM>'

Usage

import torch
from transformers import GPT2Tokenizer, T5ForConditionalGeneration 
tokenizer = GPT2Tokenizer.from_pretrained('RussianNLP/FRED-T5-Summarizer',eos_token='</s>')
model = T5ForConditionalGeneration.from_pretrained('RussianNLP/FRED-T5-Summarizer')
device='cuda'
model.to(device)

input_text='<LM> Сократи текст.\n В деревне, затерянной среди зелёных холмов, жил старик по имени Иван. Его жизнь протекала медленно и размеренно. Каждое утро Иван выходил на поля, чтобы заботиться о своём скромном участке земли. Он выращивал картофель и морковь, которые были его главным источником пищи. Вечера старик проводил у камина, читая книги и вспоминая молодость. Жизнь в деревне была тяжёлая, но Иван находил в ней простые радости.'
input_ids=torch.tensor([tokenizer.encode(input_text)]).to(device)
outputs=model.generate(input_ids,eos_token_id=tokenizer.eos_token_id,
                    num_beams=5,
                    min_new_tokens=17,
                    max_new_tokens=200,
                    do_sample=True,
                    no_repeat_ngram_size=4,
                    top_p=0.9)
print(tokenizer.decode(outputs[0][1:]))

# print result: Старик Иван живёт размеренной жизнью в деревне, выращивая овощи и находя радость в простых вещах.

Authors

  • Sber Devices:
  • Albina Akhmetgareeva
  • Ilia Kuleshov
  • Vlad Leschuk
  • Alexander Abramov
  • Alena Fenogenova

Cite us

@misc{akhmetgareeva2024summary,
      title={Towards Russian Summarization: can architecture solve data limitations problems?}, 
      author={Albina Akhmetgareeva and Ilia Kuleshov and Vlad Leschuk and Alexander Abramov and Alena Fenogenova},
      year={2024},
}

Configuration

Architecture
T5ForConditionalGeneration
Vocabulary size
50,364
Stored precision
float32
Model type
t5

Identity and Version

Repository
RussianNLP/FRED-T5-Summarizer
Publisher
Natural Language Processing in Russian
Task
Summarization
Modality
Text
Library
transformers
Parameters
1.7B parameters
Languages
ru
Revision
cb1c9278a394d91837e1b3698f847f3c0a6b5c46
First published
2024-04-02
Last updated
2024-04-22

Files and Weights

13 files, 7.0 GB in total. The weights are 3 files totalling 7.0 GB in bin, safetensors.

Weights3 files · 7.0 GB
Configuration5 files · 55.0 KB
Tokenizer3 files · 3.1 MB
Documentation1 file · 2.5 KB
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
model-00001-of-00002.safetensorsWeights5.0 GB 539b37a9498c
model-00002-of-00002.safetensorsWeights2.0 GB 0c8a5c3bca53
training_args.binWeights5.0 KB 383dc835caa1
added_tokens.jsonConfiguration2.7 KB
config.jsonConfiguration853 B
generation_config.jsonConfiguration133 B
model.safetensors.index.jsonConfiguration50.6 KB
special_tokens_map.jsonConfiguration688 B
README.mdDocumentation2.5 KB
.gitattributesRepository1.5 KB
merges.txtTokenizer1.3 MB
tokenizer_config.jsonTokenizer20.3 KB
vocab.jsonTokenizer1.8 MB

License and Download

License
mit
Access
Open weights, no gate
Download size
7.0 GB
Download from Natural Language Processing in Russian

Released by Natural Language Processing in Russian through its official repository on Hugging Face. Read the license.

Memory Requirements

PrecisionWeights in memory
As published7.0 GB
16-bit3.5 GB
8-bit1.7 GB
4-bit0.9 GB

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

Questions About FRED-T5-Summarizer

How much GPU memory does FRED-T5-Summarizer need?

About 4.2 GB at 16-bit and 1 GB at 4-bit: the weights (1.7B parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run FRED-T5-Summarizer 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 FRED-T5-Summarizer commercially?

Yes. FRED-T5-Summarizer 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.

Similar Models

Model · Summarization

mbart50-tradenewssum

Daria Lyutova

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt on the TradeNewsSum dataset for multilingual abstractive summarization of foreign trade news in Russian and English. This is a multilingual summarization model trained on economic and foreign trade news in Russian and English. It is based on the facebook/mbart-large-50-many-to-many-mmt architecture and fine-tuned specifically for the task of generating concise, informative summaries for news articles in the domain of international trade. - Generating abstractive summaries of foreign trade-related news in Russian and English. - Assisting analysts and journalists working with economic content. - Could be integrated…

Open weights mit 611M parameters 1,024 tokens transformers

This model was fine-tuned on a novel financial news dataset, which consists of 2K articles from Bloomberg, on topics such as stock, markets, currencies, rate and cryptocurrencies. It is based on the PEGASUS model and in particular PEGASUS fine-tuned on the Extreme Summarization (XSum) dataset: google/pegasus-xsum model. PEGASUS was originally proposed by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu in PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization. Note: This model serves as a base version. For an even more advanced model with significantly enhanced performance, please check out our advanced version on Rapid API. The advanced model offers more…

Open weights 569M parameters 512 tokens transformers

Model · Summarization

led-large-book-summary

Peter Szemraj

This model is a fine-tuned version of allenai/led-large-16384 on the BookSum dataset (kmfoda/booksum). It aims to generalize well and be useful in summarizing lengthy text for both academic and everyday purposes. - See the Colab demo linked above or try the demo on Spaces To improve summary quality, use encodernorepeatngramsize=3 when calling the pipeline object. This setting encourages the model to utilize new vocabulary and construct an abstractive summary. Load the model into a pipeline object: Feed the text into the pipeline object: Important: For optimal summary quality, use the global attention mask when decoding, as demonstrated in this community notebook, see the definition of…

Open weights apache-2.0 460M parameters transformers

Model · Summarization

bart-large-cnn

AI at Meta

BART model pre-trained on English language, and fine-tuned on CNN Daily Mail. It was introduced in the paper BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension by Lewis et al. and first released in this repository (https://github.com/pytorch/fairseq/tree/master/examples/bart). Disclaimer: The team releasing BART did not write a model card for this model so this model card has been written by the Hugging Face team. BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function…

Open weights mit 406M parameters 1,024 tokens transformers

Model · Summarization

distilbart-news-summarizer

Sachin Rao

Imagine you have a very long news article, and you want someone to read it and tell you the key points in just a few sentences. That's exactly what this model does! This model takes a long news article and turns it into a short, easy-to-read summary. - It writes back a 3-4 sentence summary that captures all the important information The special thing about this model is that it's: 1. Very accurate - It understands news writing style very well 2. Very fast - It works quickly even on regular computers (not just expensive AI servers) 3. Specialized in news - It was trained specifically on news articles, so it understands how journalists write 4. Good with financial news - It knows market…

Open weights agpl-3.0 306M parameters 1,024 tokens transformers