facebook / summarization updated 2 years ago

bart-large-cnn

Name: bart-large-cnn
Author: facebook

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).

Params

410 M

Context

1,024

Downloads 30d

1.6 M

Likes

1,606

Commercial use: allowed mit Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

1.6 M1.5 M

Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	f32	1.6 GB	2.3 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Architecture: BartForConditionalGeneration
Parameters: 410 M
Tensor type: F32
Context length: 1,024
Vocabulary: 50,264
Licence: mit
First seen on the Hub: 2022-03-02
Training datasets: cnn_dailymail
cnn_dailymail (reported): 78.5866
Added to our catalog: 2026-07-28

Compare with

distilbart-cnn-12-6 851 K pegasus-xsum 187 K financial-summarization-pegasus 153 K bart-large-cnn-samsum 130 K T5-base-10K-summarization 84 K distilbart-xsum-12-6 76 K

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB