mt5-small-prompted-germanquad-1

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

mt5-small-prompted-germanquad-1

This model is a fine-tuned version of google/mt5-small on an philschmid/prompted-germanquad dataset. A prompt datasets using the BigScience PromptSource library. The dataset is a copy of germanquad with applying the squad template and translated it to german. TEMPLATE.

This is a first test if it is possible to fine-tune mt5 models to solve similar tasks than T0 of big science but for the German language.

It achieves the following results on the evaluation set:

• Loss: 1.6835
• Rouge1: 27.7309
• Rouge2: 18.7311
• Rougel: 27.4704
• Rougelsum: 27.4818

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5.6e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 7

Training results

Framework versions

• Transformers 4.14.1
• Pytorch 1.10.1+cu102
• Datasets 1.16.1
• Tokenizers 0.10.3