metadata
license: apache-2.0
base_model: tsmatz/mt5_summarize_japanese
tags:
- generated_from_trainer
datasets:
- xlsum
metrics:
- rouge
model-index:
- name: mt5_summarize_japanese-6051-japanese
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: xlsum
type: xlsum
config: japanese
split: validation
args: japanese
metrics:
- name: Rouge1
type: rouge
value: 0.3394
mt5_summarize_japanese-6051-japanese
This model is a fine-tuned version of tsmatz/mt5_summarize_japanese on the xlsum dataset. It achieves the following results on the evaluation set:
- Loss: 1.1045
- Rouge1: 0.3394
- Rouge2: 0.0501
- Rougel: 0.3328
- Rougelsum: 0.3321
- Gen Len: 31.8808
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.3889 | 4.5 | 500 | 1.1112 | 0.3381 | 0.05 | 0.3306 | 0.3302 | 31.3127 |
| 1.3351 | 8.99 | 1000 | 1.1045 | 0.3394 | 0.0501 | 0.3328 | 0.3321 | 31.8808 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0