metadata
language:
- de
library_name: transformers
base_model: google/mT5-base
datasets:
- DEplain/DEplain-APA-sent
metrics:
- sari
- bleu
- bertscore
pipeline_tag: text2text-generation
Model Card for mT5-base-trimmed_deplain-apa
Finetuned mT5-Model for German sentence-level text-simplification.
Model Details
Model Description
- Model type: Encoder-Decoder-Transformer
- Language(s) (NLP): German
- Finetuned from model: google/mT5-base
- Task: Text-Simplification
Training Details
Training Data
DEplain/DEplain-APA-sent
Stodden et al. (2023):arXiv:2305.18939
Training Procedure
Parameter-efficient Fine-Tuning with LoRA. Vocabulary trimmed to 32.000 most frequent tokens for German.
Training Hyperparameters
- Batch Size: 16
- Epochs: 1
- Learning Rate: 0.001
- Optimizer: Adafactor
LoRA Hyperparameters
- R: 32
- Alpha: 64
- Dropout: 0.1
- Target modules: all linear layers