|
|
--- |
|
|
license: apache-2.0 |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
metrics: |
|
|
- bleu |
|
|
model-index: |
|
|
- name: HingFlow |
|
|
results: [] |
|
|
datasets: |
|
|
- cfilt/iitb-english-hindi |
|
|
language: |
|
|
- en |
|
|
library_name: transformers |
|
|
pipeline_tag: translation |
|
|
--- |
|
|
|
|
|
# HingFlow |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.1887 |
|
|
- Bleu: 72.3468 |
|
|
- Gen Len: 5.9953 |
|
|
|
|
|
## Model description |
|
|
https://github.com/SRDdev/HingFlow |
|
|
|
|
|
## 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: 2 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |
|
|
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:| |
|
|
| 0.1505 | 1.0 | 1000 | 0.2053 | 71.6108 | 5.8418 | |
|
|
| 0.1057 | 2.0 | 2000 | 0.1887 | 72.3468 | 5.9953 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.29.0 |
|
|
- Pytorch 2.0.0+cu118 |
|
|
- Datasets 2.12.0 |
|
|
- Tokenizers 0.13.3 |
