|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- wmt19 |
|
metrics: |
|
- bleu |
|
model-index: |
|
- name: t5-small-finetuned-fi-to-en |
|
results: |
|
- task: |
|
name: Sequence-to-sequence Language Modeling |
|
type: text2text-generation |
|
dataset: |
|
name: wmt19 |
|
type: wmt19 |
|
args: fi-en |
|
metrics: |
|
- name: Bleu |
|
type: bleu |
|
value: 1.618 |
|
--- |
|
|
|
<!-- 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-small-finetuned-fi-to-en |
|
|
|
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the wmt19 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 3.3598 |
|
- Bleu: 1.618 |
|
- Gen Len: 17.3223 |
|
|
|
## 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: 2e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 5 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:|:-------:| |
|
| 3.3627 | 1.0 | 6250 | 3.5122 | 1.2882 | 17.1803 | |
|
| 3.2162 | 2.0 | 12500 | 3.4442 | 1.4329 | 17.2617 | |
|
| 3.1304 | 3.0 | 18750 | 3.3872 | 1.4862 | 17.296 | |
|
| 3.0832 | 4.0 | 25000 | 3.3648 | 1.5795 | 17.3047 | |
|
| 3.0623 | 5.0 | 31250 | 3.3598 | 1.618 | 17.3223 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.12.5 |
|
- Pytorch 1.9.1 |
|
- Datasets 1.16.1 |
|
- Tokenizers 0.10.3 |
|
|