metadata
base_model: facebook/mbart-large-50
library_name: peft
license: mit
tags:
- generated_from_trainer
model-index:
- name: mbart-large-50_Nepali_News_Summarization_0
results: []
datasets:
- caspro/Nepali_News_Dataset
language:
- ne
metrics:
- rouge
pipeline_tag: text2text-generation
mbart-large-50_Nepali_News_Summarization_0
This model is a fine-tuned version of facebook/mbart-large-50 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.3697
- Rouge-1 R: 0.3809
- Rouge-1 P: 0.3915
- Rouge-1 F: 0.3761
- Rouge-2 R: 0.215
- Rouge-2 P: 0.2209
- Rouge-2 F: 0.2109
- Rouge-l R: 0.3708
- Rouge-l P: 0.3809
- Rouge-l F: 0.3661
- Gen Len: 13.9732
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: 0.0005
- train_batch_size: 5
- eval_batch_size: 5
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge-1 R | Rouge-1 P | Rouge-1 F | Rouge-2 R | Rouge-2 P | Rouge-2 F | Rouge-l R | Rouge-l P | Rouge-l F | Gen Len |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.563 | 1.0 | 10191 | 1.4569 | 0.3599 | 0.3789 | 0.3581 | 0.1965 | 0.2082 | 0.1946 | 0.3499 | 0.3683 | 0.3481 | 13.8415 |
| 1.4053 | 2.0 | 20382 | 1.3963 | 0.3631 | 0.3927 | 0.367 | 0.203 | 0.2203 | 0.204 | 0.354 | 0.3827 | 0.3577 | 13.4601 |
| 1.2399 | 3.0 | 30573 | 1.3697 | 0.3809 | 0.3915 | 0.3761 | 0.215 | 0.2209 | 0.2109 | 0.3708 | 0.3809 | 0.3661 | 13.9732 |
Framework versions
- PEFT 0.11.1
- Transformers 4.42.3
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1