mt5-small-german-query-generation

Model description:

This model was created with the purpose to generate possible queries for a german input article.

For this model, we finetuned a multilingual T5 model mt5-small on the MMARCO dataset the machine translated version of the MS MARCO dataset.

The model was trained for 1 epoch, on 200,000 unique queries of the dataset. We trained the model on one K80 GPU for 25,000 iterations with following parameters:

• learning rate: 1e-3
• train batch size: 8
• max input sequence length: 512
• max target sequence length: 64

Model Performance:

Model evaluation was done on 2000 evaluation paragraphs of the dataset. Mean f1 ROUGE scores were calculated for the model.

Rouge-1	Rouge-2	Rouge-L
0.162	0.052	0.161