This Model is a fine-tuned version of T-systems
summarization model v1.
We used 1000 examples of headline-content pairs from BR24 articles for the fine-tuning process.
Despite the small amount of training data, the tonality of the summarizations has changed significantly. Many of the resulting summaries do sound like a headline.
Since the base model is trained on the MLSUM dataset, this model may not be used for commercial use.
