# Which CNN Summarization models to use?
**URL:**
**Author:**  [@echatzikyriakidis](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fechatzikyriakidis)%5C
**Post date:** [July 16, 2020, 9:05am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F1 "2020-07-16T09:05:36Z")
Hi @sshleifer,
For what task the following models were fine tuned / trained? Can be used for text summarization?
sshleifer/student\_cnn\_12\_6
sshleifer/student\_cnn\_6\_6
Thank you.
---
**Author:**  [@valhalla](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fvalhalla)%5C
**Post date:** [July 16, 2020, 2:13pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F2 "2020-07-16T14:13:18Z")
These student models are created by copying layers from bart-large-cnn to reduce their size. These are un fine-tuned checkpoints so you’ll need to fine-tune them for summerization. More details can be found here
[https://github.com/huggingface/transformers/tree/master/examples/seq2seq#distilbart](https://github.com/huggingface/transformers/tree/master/examples/seq2seq#distilbart)
---
**Author:**  [@echatzikyriakidis](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fechatzikyriakidis)%5C
**Post date:** [July 16, 2020, 2:50pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F3 "2020-07-16T14:50:16Z")
Thank you!
---
**Author:**  [@sshleifer](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsshleifer)%5C
**Post date:** [July 16, 2020, 4:55pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F4 "2020-07-16T16:55:16Z")
Yes @valhalla is 100% correct.
More info:
Those are the starting point for distillation experiments
that result in distilbart [https://docs.google.com/spreadsheets/d/1EkhDMwVO02m8jCD1cG3RoFPLicpcL1GQHTQjfvDYgIM/edit?usp=sharing](https://docs.google.com/spreadsheets/d/1EkhDMwVO02m8jCD1cG3RoFPLicpcL1GQHTQjfvDYgIM/edit?usp=sharing)
---
**Author:**  [@echatzikyriakidis](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fechatzikyriakidis)%5C
**Post date:** [July 17, 2020, 8:55am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F5 "2020-07-17T08:55:39Z")
Thank you @sshleifer!
---
**Author:**  [@echatzikyriakidis](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fechatzikyriakidis)%5C
**Post date:** [July 17, 2020, 9:46am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F6 "2020-07-17T09:46:07Z")
Hi Sam, @sshleifer
I would like to ask if I can find also such metrics for other models
e.g., airKlizz/_, t5-_ and mrm8488/t5-base-finetuned-summarize-news
Thank you.
---
**Author:**  [@valhalla](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fvalhalla)%5C
**Post date:** [July 17, 2020, 10:31am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F7 "2020-07-17T10:31:20Z")
cc @mrm8488
---
**Author:**  [@mrm8488](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fmrm8488)%5C
**Post date:** [July 17, 2020, 10:44am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F8 "2020-07-17T10:44:11Z")
Add the metrics is in my TODO lists
---
**Author:**  [@sshleifer](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsshleifer)%5C
**Post date:** [July 17, 2020, 3:09pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F9 "2020-07-17T15:09:33Z")
Yes. It is up to whoever uploaded the model to post their metrics. Please use rouge scores for summarization. Ideally use the `nlp` package (`nlp.metrics('rouge')` or the `calculate_rouge_score` function so that we can compare apples to apples, and make sure that beam search params are in your config!
Metrics that matter the most:
Rouge2 (fscore mid measure)
RougeL (fscore mid measure)
---
**Author:**  [@echatzikyriakidis](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fechatzikyriakidis)%5C
**Post date:** [November 26, 2020, 12:33pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F10 "2020-11-26T12:33:49Z")
Dear all!
Which of the following models can be used for summarizating English text out-of-the-box? Without the need to fine-tune any of them.
Thank you!
Here are the models:
google/pegasus-xsum
google/pegasus-cnn\_dailymail
google/pegasus-large
google/pegasus-multi\_news
google/pegasus-arxiv
google/pegasus-wikihow
google/pegasus-gigaword
google/pegasus-pubmed
google/pegasus-newsroom
google/pegasus-reddit\_tifu
google/pegasus-billsum
google/pegasus-aeslc
google/roberta2roberta\_L-24\_bbc
google/roberta2roberta\_L-24\_gigaword
google/roberta2roberta\_L-24\_cnn\_daily\_mail
sshleifer/distill-pegasus-cnn-16-4
sshleifer/distill-pegasus-xsum-16-8
sshleifer/distill-pegasus-xsum-16-4
sshleifer/pegasus-cnn-ft-v2
patrickvonplaten/roberta\_shared\_bbc\_xsum
patrickvonplaten/bert2bert\_cnn\_daily\_mail
mrm8488/bert-small2bert-small-finetuned-cnn\_daily\_mail-summarization
yuvraj/xSumm
yuvraj/summarizer-cnndm
---
**Author:**  [@valhalla](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fvalhalla)%5C
**Post date:** [December 11, 2020, 10:15am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fwhich-cnn-summarization-models-to-use%2F317%2F11 "2020-12-11T10:15:59Z")
All of these models are fine-tuned for summarization, you could select the model based on your requirements like domain, model size, extractive/abstractive etc.