|
--- |
|
license: apache-2.0 |
|
base_model: facebook/bart-large |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- rouge |
|
- precision |
|
- recall |
|
- f1 |
|
model-index: |
|
- name: LLM_Teached_Bart_100k |
|
results: [] |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# LLM_Teached_Bart_100k |
|
|
|
This model is a fine-tuned version of [facebook/bart-large](https://huggingface.co/facebook/bart-large) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.4623 |
|
- Rouge1: 0.4678 |
|
- Rouge2: 0.2472 |
|
- Rougel: 0.4081 |
|
- Rougelsum: 0.4082 |
|
- Gen Len: 19.8816 |
|
- Precision: 0.9185 |
|
- Recall: 0.8957 |
|
- F1: 0.9068 |
|
|
|
## 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: 24 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 96 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 16 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|:---------:|:------:|:------:| |
|
| 1.6328 | 1.0 | 1041 | 1.4801 | 0.448 | 0.2243 | 0.385 | 0.385 | 19.8782 | 0.9134 | 0.893 | 0.9029 | |
|
| 1.4598 | 2.0 | 2083 | 1.4051 | 0.4428 | 0.2273 | 0.3851 | 0.385 | 19.9344 | 0.9147 | 0.8903 | 0.9022 | |
|
| 1.3402 | 3.0 | 3125 | 1.3840 | 0.4498 | 0.2318 | 0.3921 | 0.392 | 19.95 | 0.9158 | 0.8918 | 0.9034 | |
|
| 1.2446 | 4.0 | 4167 | 1.3682 | 0.4604 | 0.2405 | 0.4014 | 0.4014 | 19.884 | 0.9169 | 0.8944 | 0.9054 | |
|
| 1.1651 | 5.0 | 5208 | 1.3695 | 0.4594 | 0.2401 | 0.3995 | 0.3995 | 19.894 | 0.9173 | 0.8942 | 0.9055 | |
|
| 1.1002 | 6.0 | 6250 | 1.3783 | 0.4607 | 0.2423 | 0.4014 | 0.4014 | 19.9118 | 0.9166 | 0.8945 | 0.9053 | |
|
| 1.0427 | 7.0 | 7292 | 1.3851 | 0.462 | 0.2432 | 0.4028 | 0.4028 | 19.9075 | 0.9172 | 0.8946 | 0.9056 | |
|
| 0.9881 | 8.0 | 8334 | 1.3911 | 0.4635 | 0.2442 | 0.4038 | 0.4037 | 19.9071 | 0.9177 | 0.8947 | 0.9059 | |
|
| 0.9435 | 9.0 | 9375 | 1.4075 | 0.468 | 0.2471 | 0.4085 | 0.4084 | 19.8805 | 0.918 | 0.8959 | 0.9067 | |
|
| 0.9035 | 10.0 | 10417 | 1.4125 | 0.4675 | 0.248 | 0.4085 | 0.4086 | 19.8811 | 0.9178 | 0.8957 | 0.9064 | |
|
| 0.8702 | 11.0 | 11459 | 1.4219 | 0.4646 | 0.2455 | 0.405 | 0.4051 | 19.8947 | 0.9181 | 0.895 | 0.9063 | |
|
| 0.8458 | 12.0 | 12501 | 1.4339 | 0.4643 | 0.2447 | 0.4055 | 0.4055 | 19.8985 | 0.9177 | 0.8952 | 0.9061 | |
|
| 0.8207 | 13.0 | 13542 | 1.4430 | 0.4671 | 0.2463 | 0.4068 | 0.4069 | 19.9053 | 0.9182 | 0.8952 | 0.9064 | |
|
| 0.7987 | 14.0 | 14584 | 1.4495 | 0.4633 | 0.2455 | 0.4046 | 0.4047 | 19.918 | 0.9179 | 0.8944 | 0.9059 | |
|
| 0.787 | 15.0 | 15626 | 1.4560 | 0.4666 | 0.2471 | 0.407 | 0.4072 | 19.8956 | 0.9182 | 0.8953 | 0.9064 | |
|
| 0.772 | 15.99 | 16656 | 1.4623 | 0.4678 | 0.2472 | 0.4081 | 0.4082 | 19.8816 | 0.9185 | 0.8957 | 0.9068 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.36.0 |
|
- Pytorch 2.0.1+cu117 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.15.0 |
|
|
