Distilled 12-6 configuration
According to the model card's metrics table, this configuration lists 306M parameters with 137ms inference time and a 1.68x speedup.
Open Source Model Profile · sshleifer
distilbart-xsum-12-6
distilbart-xsum-12-6 is a distilled BART summarization model from sshleifer. According to the model card, it reports ROUGE-2 22.12 and ROUGE-L 36.99 with a 1.68x speedup over the BART-large baseline.
- Publisher
- sshleifer
- Task
- summarization
- Model type
- bart
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
distilbart-xsum-12-6 is published by sshleifer for summarization in the XSum style. The captured configuration identifies BartForConditionalGeneration with model type bart, and hub tags record cnn_dailymail and xsum datasets. According to the model card, the checkpoint loads through BartForConditionalGeneration.from_pretrained, and the metrics table lists 306M parameters with a 1.68x speedup over baseline.
Recorded capabilities
Reported ROUGE scores
According to the model card, the model reports ROUGE-2 22.12 and ROUGE-L 36.99, against the listed BART-large baseline of 21.85 and 36.50.
XSum and CNN/DM dataset tags
Hub tags record cnn_dailymail and xsum dataset associations with English summarization support.
Documented load path
According to the model card, the checkpoint loads into BartForConditionalGeneration.from_pretrained with further detail in the BART documentation.
Use cases in the source record
- Extreme summarization in the XSum style through Transformers sequence-to-sequence workflows.
- Latency-sensitive summarization where the publisher's table reports about 137ms inference time with a 1.68x speedup over the BART-large baseline.
Limitations and unknowns
- No Safetensors parameter count was captured; the 306M figure comes from the publisher's metrics table only.
- No context-window value was extracted from this record.
- Speed and ROUGE figures are publisher-reported table values and were not independently verified by Ethen.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
Source and provenance
Source: sshleifer/distilbart-xsum-12-6
Captured: Unknown. Processed: 2026-09-07T19:34:58.765619+00:00.
Usage This checkpoint should be loaded into BartForConditionalGeneration.from_pretrained . See the BART docs for more information. Metrics for DistilBART models Model Name MM Params Inference Time (MS) Speedup Rouge 2 Rouge-L distilbart-xsum-12-1 222 90 2.54 18.31 33.37 distilbart-xsum-6-6 230 132 1.73 20.92 35.73 distilbart-xsum-12-3 255 106 2.16 21.37 36.39 distilbart-xsum-9-6 268 136 1.68 21.72 36.61 bart-large-xsum (baseline) 406 229 1 21.85 36.50 distilbart-xsum-12-6 306 137 1.68 22.12 36.99 bart-large-cnn (baseline) 406 381 1 21.06 30.63 distilbart-12-3-cnn 255 214 1.78 20.57 30.00 distilbart-12-6-cnn 306 307 1.24 21.26 30.59…
F001F002F003F004F005F006F008F009F010F011