|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- esnli |
|
metrics: |
|
- accuracy |
|
- f1 |
|
- rouge |
|
- bleu |
|
model-index: |
|
- name: google-flan-t5-small-e-snli-generation-label_and_explanation-selected-b48 |
|
results: |
|
- task: |
|
name: Sequence-to-sequence Language Modeling |
|
type: text2text-generation |
|
dataset: |
|
name: esnli |
|
type: esnli |
|
config: plain_text |
|
split: validation |
|
args: plain_text |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 0.8622231253810201 |
|
- name: F1 |
|
type: f1 |
|
value: 0.8623314280769628 |
|
- name: Rouge1 |
|
type: rouge |
|
value: 0.605873896307076 |
|
- name: Bleu |
|
type: bleu |
|
value: 0.40472213589689604 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# google-flan-t5-small-e-snli-generation-label_and_explanation-selected-b48 |
|
|
|
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co./google/flan-t5-small) on the esnli dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.8720 |
|
- Accuracy: 0.8622 |
|
- F1: 0.8623 |
|
- Bertscore F1: 0.9329 |
|
- Rouge1: 0.6059 |
|
- Rouge2: 0.3988 |
|
- Rougel: 0.5475 |
|
- Rougelsum: 0.5496 |
|
- Bleu: 0.4047 |
|
|
|
## 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: 0.001 |
|
- train_batch_size: 48 |
|
- eval_batch_size: 48 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.05 |
|
- num_epochs: 10 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Bertscore F1 | Rouge1 | Rouge2 | Rougel | Rougelsum | Bleu | |
|
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:------:|:------------:|:------:|:------:|:------:|:---------:|:------:| |
|
| 1.5084 | 0.17 | 2000 | 1.7484 | 0.8001 | 0.7997 | 0.9271 | 0.5768 | 0.3695 | 0.5209 | 0.5229 | 0.3703 | |
|
| 1.2745 | 0.35 | 4000 | 1.8137 | 0.8113 | 0.8110 | 0.9304 | 0.5881 | 0.3804 | 0.5305 | 0.5325 | 0.3853 | |
|
| 1.2287 | 0.52 | 6000 | 1.8358 | 0.8392 | 0.8403 | 0.9298 | 0.5828 | 0.3747 | 0.5282 | 0.5301 | 0.3778 | |
|
| 1.1964 | 0.7 | 8000 | 1.8432 | 0.8430 | 0.8437 | 0.9326 | 0.5974 | 0.3905 | 0.5447 | 0.5462 | 0.3998 | |
|
| 1.1674 | 0.87 | 10000 | 1.8567 | 0.8507 | 0.8485 | 0.9310 | 0.5947 | 0.3888 | 0.5383 | 0.5402 | 0.3892 | |
|
| 1.1371 | 1.05 | 12000 | 1.8720 | 0.8622 | 0.8623 | 0.9329 | 0.6059 | 0.3988 | 0.5475 | 0.5496 | 0.4047 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.27.4 |
|
- Pytorch 2.0.0+cu117 |
|
- Datasets 2.11.0 |
|
- Tokenizers 0.13.2 |
|
|
