Downloads · 30 days
13
24% of all-time downloads
kobigbird-bert-base-finetuned-klue-v2
Seongmi/kobigbird-bert-base-finetuned-klue-v2
kobigbird-bert-base-finetuned-klue-v2 is a question answering model from Seongmi. Use it when the input is a question plus a passage. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
54
Public
Repo size
2.4 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin472 MB · 100%
At a glance
- Task
- Question Answering
- Library
- transformers
- Model type
- big_bird
- Access
- Public
- Created
- Oct 7, 2022
- Updated
- Oct 7, 2022
- SHA
- f455eadc
Try a prompt
Model card
From the Hugging Face model README
<!-- 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. -->
kobigbird-bert-base-finetuned-klue-v2
This model is a fine-tuned version of monologg/kobigbird-bert-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9250
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: 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: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.5274 | 0.35 | 500 | 2.7400 |
| 1.9296 | 0.69 | 1000 | 1.4055 |
| 1.3583 | 1.04 | 1500 | 1.1407 |
| 1.1119 | 1.38 | 2000 | 1.0246 |
| 1.017 | 1.73 | 2500 | 0.9231 |
| 0.8637 | 2.07 | 3000 | 0.9507 |
| 0.6573 | 2.42 | 3500 | 0.9678 |
| 0.6916 | 2.76 | 4000 | 0.9250 |
Framework versions
- Transformers 4.22.2
- Pytorch 1.12.1+cu113
- Datasets 2.5.2
- Tokenizers 0.12.1
- Task
- Question Answering
- Library
- transformers
- Type
- big_bird
- Created
- Oct 7, 2022
- Updated
- Oct 7, 2022