language: | |
- de | |
tags: | |
- generated_from_trainer | |
metrics: | |
- rouge | |
model-index: | |
- name: T5-Base_GNAD_MaxSamples | |
results: [] | |
<!-- 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. --> | |
# T5-Base_GNAD_MaxSamples | |
This model is a fine-tuned version of [Einmalumdiewelt/T5-Base_GNAD_MaxSamples](https://huggingface.co/Einmalumdiewelt/T5-Base_GNAD_MaxSamples) on an unknown dataset. | |
It achieves the following results on the evaluation set: | |
- Loss: 2.0228 | |
- Rouge1: 28.4344 | |
- Rouge2: 9.3461 | |
- Rougel: 19.7157 | |
- Rougelsum: 24.6508 | |
- Gen Len: 54.4873 | |
## 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: 5e-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 | |
- num_epochs: 10.0 | |
### Training results | |
### Framework versions | |
- Transformers 4.22.0.dev0 | |
- Pytorch 1.12.1+cu113 | |
- Datasets 2.4.0 | |
- Tokenizers 0.12.1 | |